Haggerty, J.A., Premoli Silva, I., Rack, F., and McNutt, M.K. (Eds.), 1995 Proceedings of the Ocean Drilling Program, Scientific Results, Vol. 144
30. MINERAL COMPOSITIONS IN LEG 144 LAVAS AND ULTRAMAFIC XENOLITHS: THE ROLES OF CUMULATES AND CARBONATITE METASOMATISM IN MAGMA PETROGENESIS1
Julie J. Dieu2
ABSTRACT
Severe alteration of many of the Cretaceous-aged hotspot lavas recovered by drilling during Leg 144 of the Ocean Drilling Program made geochemical analysis of minerals an attractive alternative to whole-rock analysis. Igneous rock samples were recovered from Limalok, Lo-En, Wodejebato, MIT, and Takuyo-Daisan guyots. Unaltered groundmass grains, microlites, and phenocrysts of olivine, plagioclase, and clinopyroxene in representative lavas were analyzed by electron microprobe for major elements, Cr, and Ni. These data show that the lavas recovered during Leg 144 are moderately to very silica-undersaturated and belong to the alkalic series or, in some cases, to the basanitic series as identified in Hawaii; exceptions are lavas from Lo-En and Takuyo-Daisan guyots, which are more transitional in nature. Ion microprobe analyses for rare earth elements, Sr, Cr, Y, V, Ti, and Zr on selected microlites and phenocrysts of clinopyroxene show that (1) microphenocrysts contain considerably higher abundances of incompatible elements than phenocrysts from the same lava because of syn- and post-eruption fractionation; (2) the basanitic lavas from Limalok Guyot sampled a different, and probably more enriched, source material than the basanitic lavas from MIT Guyot; (3) the transitional and alkalic series lavas from Lo-En, MIT, and Takuyo-Daisan guyots contain phenocrysts with similar concentrations of incompatible trace elements, despite differ- ences in silica-saturation and source composition; and (4) relative to other lavas, most of the Wodejebato Guyot alkalic series lavas contain phenocrysts with higher trace element concentrations and higher La/Yb, probably because of differences in depth of crystallization and/or source composition. Ultramafic xenoliths were recovered in lavas from Limalok and MIT guyots. Electron probe analyses of clinopyroxene and orthopyroxene demonstrate that both Type 1 (mantle residue) and Type 2 (cumulate) xenoliths were recovered from each guyot. The Type 2 xenoliths from Limalok Guyot are the cumulate products of differentiation of alkalic liquids. Those from MIT Guyot are the cumulate products of fractionation of tholeiitic liquids. Ion microprobe analyses show that carbonatite metasomatism of mantle residue (probably oceanic lifhosphere) by a plume process observed throughout the SOPITA region and in the Kerguelen Islands is not only ubiquitous in space across the Southern Hemisphere, but ubiquitous in time as early as the mid-Cretaceous (MIT Guyot is 128 Ma).
INTRODUCTION Igneous rock was recovered from nine sites on five guyots drilled during Leg 144. Drilling penetrated 48.5 m of basanitic lava flows and Ocean Drilling Program (ODP) Leg 144 was the second "atolls and volcaniclastic breccias in Hole 871C on Limalok Guyot and 57.1 m of guyots" leg, sister to Leg 143. The objective was to drill Limalok, differentiated transitional lava flows in Hole 872B on Lo-En Guyot. Lo-En, and Wodejebato guyots in the Marshall Seamount Chain, as Drilling at each of five sites on Wodejebato Guyot recovered primitive well as MIT Guyot (an isolated feature far northwest of the Marshall to somewhat fractionated alkalic series basalts. In Hole 873A, volcanic Islands) and Takuyo-Daisan Guyot in the Japanese Seamount Group. debris-flow breccias and 29.2 m of basalt were recovered. In Hole These guyots are drowned atolls built on submerging volcanic islands 874B, 15.9 m of ankaramitic basalt were penetrated. In Hole 875C, 6.7 formed during the Cretaceous in the southern Pacific Ocean. The prin- m of highly altered and fractured basalt were drilled. In Hole 876A, 9.0 ciple objectives for recovery of igneous basement were to compare m of basalt were penetrated. In Hole 877 A, 1.0 m of altered basalt was lava types and igneous geochemistry of the drilled guyots and to com- recovered. On MIT Guyot (Hole 878A), drilling penetrated 187.5 m of pare the igneous geochemistry of lavas recovered from these features "basement" rocks. The recovered igneous rocks include volcanogenic with modern lavas in the South Pacific Isotopic and Thermal Anomaly sandstone, lithic tuff, basalt breccia, and alkalic and basanitic series (SOPITA) (Staudigel et al, 1991) or Dupal region (Hart, 1984). lavas. On Takuyo-Daisan Guyot (Hole 879A), drilling penetrated 34.4 The SOPITA region is an area of considerable hotspot volcanism, m of a complex, intercalated mixture of clinopyroxene-plagioclase and at least five hotspots are currently or recently active in the SOPITA: phyric basalt and volcanic breccia, which appears to be a peperite Marquesan, Society, Samoan, Pitcairn, and Cook-Austral (see Koppers (intrusion of lava into soft, wet sediment). et al., this volume, for a location map). Basalts erupted at these hotspots Some of these lavas are relatively unaltered, but many are altered display unusual, and often extreme, isotopic and incompatible element to claystone (which retains relict igneous texture), and their whole- ratios and incompatible element concentrations when compared with rock compositions do not reflect magmatic compositions. However, lavas erupted at other hotspots around the globe (e.g., Vidal et al., 1984; unaltered clinopyroxene is ubiquitous and unaltered plagioclase is Wright and White, 1986/1987; Staudigel et al., 1991; Chauvel et al., common. The major element compositions of these phases can be used 1992). The SOPITA appears to have been active since the Cretaceous as to characterize, evaluate, and compare all of these lavas, regardless of short-lived volcanic chains with lavas geochemically similar to SOPITA the extent of whole-rock alteration. In some cases, the phenocryst lavas cover much of the western Pacific basin (Staudigel et al., 1991). compositions support the conclusions derived from petrographic and whole-rock geochemical analyses. But in other cases, evaluation of phenocryst compositions provide the definitive characterization of Haggerty, J.A., Premoli Silva, I., Rack, E, and McNutt, M.K. (Eds.), 1995. Proc. a lava or a suite of lavas. For example, groundmass clinopyroxene ODP, Sci. Results, 144: College Station, TX (Ocean Drilling Program). compositions demonstrate that the Lo-En Guyot lavas are transitional 2 Scripps Institution of Oceanography, University of California at San Diego, Mail Code 0208, San Diego, CA 92093, U.S.A. (Present address: Rayonier, P.O. Box 200, rather than truly alkalic, as petrographic and whole-rock major ele- Hoquiam, WA 98550, U.S.A.) ment analyses have indicated.
513 J J. DIEU
Trace element concentrations and ratios in clinopyroxene can basalt of MIT Guyot (igneous Unit 15, Hole 878A). Forsterite con- demonstrate fundamental differences in mantle sources of lavas be- tents of unaltered olivine phenocrysts from these lavas range from cause trace element concentrations in clinopyroxene have some di- 81% to 86% (Table 2), corresponding to liquid Mg#s of 0.62 to 0.65 rect, although not simple, relationships to the trace element concen- (calculations after Roeder and Emslie, 1970). Liquids as magnesian trations in the lava from which it crystallizes. These differences may as these are not typical of Leg 144 lavas. Specifically, the clinopy- be difficult to establish from the whole-rock geochemistry because of roxene compositions of the fresh-olivine-bearing basalts are more the effects of alteration. Furthermore, trace element analyses of care- primitive than the majority of Leg 144 lavas. fully chosen clinopyroxene grains can be used to evaluate the extent of syn- and post-eruption differentiation of a liquid. The first-order objective of the research I present in this paper is Clinopyroxene to compare the geochemistry of Cretaceous-aged, western Pacific hotspot lavas with SOPITAand Hawaiian lavas through an evaluation The interpretation of magma characteristics from clinopyroxene of their mineral chemistry. Because distinct geochemical differences compositions is a complex task that requires distinguishing the effects are seen between SOPITA and other hotspot lavas (Hawaii is chosen of the pressure of crystallization, magma chemistry, and magma as a Northern Hemisphere baseline because so much data is avail- temperature, all of which exert strong controls on clinopyroxene able), these differences should be apparent in the mineral chemistry. chemistry (e.g., Le Bas, 1962; Mori and Green, 1976; Wells, 1977; If these differences in mineral chemistry can be established, the Herzberg, 1978; Gill, 1981; Davies et al, 1989). A further complica- relationships of even very altered western Pacific lavas to SOPITA tion arises because most clinopyroxene compositions used in evalu- lavas can be evaluated. However, at this time very few trace element ating the nature of the crystallizing liquid were wet chemical analyses data of phenocrysts from either Hawaiian lavas or SOPITA lavas of mineral separates (e.g., Le Bas, 1962, figs. 1,2,5). A wet chemical exist. Thus, publication of the data presented in this paper is only the analysis probably represents the average clinopyroxene composition first step toward this larger objective. crystallized from a liquid. When electron microprobe analyses are Ultramafic xenoliths were recovered in lavas from Limalok and evaluated, more scatter is expected because both clinopyroxene phe- MIT guyots. Chemical analyses of xenoliths provide information nocryst and groundmass compositions from a single lava vary widely about the underlying lithosphere, about plume chemistry and pro- when analyzed in the microscopic detail made possible by the elec- cesses, and about magmas that cannot be sampled at the surface. The tron microprobe. silicate phases of xenoliths recovered by drilling during Leg 144 were I use three diagrams to characterize the Leg 144 lavas from major analyzed for major and trace element concentrations. These results element chemistry of clinopyroxene. A plot of TiO2 vs. Mg# (Fig. 1) are discussed in this paper ("Ultramafic Xenoliths"), along with the shows variations in the relative degree of differentiation among lavas results of analyses of phenocrysts, microphenocrysts, and microlites and may provide some insight into the tholeiitic or alkalic nature. ("Cogenetic Minerals"). However, there is considerable overlap in the clinopyroxene compo- sitions from Hawaiian tholeiitic and alkalic lavas (Fig. 1 A). By itself, this diagram is not diagnostic of the silica-saturation of a magma. ANALYTICAL TECHNIQUES The second diagram is a plot of Al(z) vs. TiO2 after Le Bas (1962) IV Compositions of silicate minerals for major elements, Ni, and Cr (Fig. 2). [Al(z) = A1 × 100/2; specifically, Al(z) is the percentage of were obtained using a Cameca CAMEB AX Microbeam electron probe tetrahedral sites occupied by AL] The nonalkaline/alkaline and alka- at Scripps Institution of Oceanography (SIO). Analyses were per- line/peralkaline boundaries on this diagram were determined from formed on polished 30-µm thin sections. Thin sections made aboard groundmass compositions (Le Bas, 1962). In general, the Le Bas ship were repolished at SIO, and the author's shore-based thin sections (1962) nonalkaline and alkaline categories roughly correspond to were made at SIO (Table 1). All of the latter are 1 in. diameter rounds tholeiitic and alkalic as defined for hotspot lavas. But characterization because this was the only size that could be introduced into the ion and interpretation of the Leg 144 lavas are facilitated by plotting both microprobe. Data reduction included routine atomic number, absorp- phenocryst and groundmass compositions. Figure 2 shows the fields tion, and fluorescence (ZAF) corrections. However, analyses were of Hawaiian tholeiitic and alkalic phenocryst compositions to assist further normalized to analyses of standard minerals (USNM Kakanui interpretation of Leg 144 phenocryst data. Two effects from magma Augite, USNM San Carlos Olivine, USNM Lake County Plagioclase) chemistry are observed on this plot. First, the silica-saturation of a obtained as unknowns. Selected analyses of olivine, clinopyroxene, magma controls the amount of tetrahedral aluminum substitution. Titanium positively correlates with Al(z) because of the cation charge and feldspar minerals appear in Tables 2-4, respectively. All of the 3+ 4+ electron probe data are on the CD-ROM (back pocket). deficit in the tetrahedral sites (Al vs. Si ). Second, differentiation drives clinopyroxene compositions to lower Al(z) and higher Ti for Ion microprobe analyses of clinopyroxene grains were made at tholeiitic liquids and to higher Al(z) and higher Ti for alkalic liquids Woods Hole Oceanographic Institution on a Cameca 3f. Rare earth (Fig. 2B). As with Figure 1, this diagram is not uniquely diagnostic elements analyzed were La, Ce, Nd, Sm, Eu, Dy, Er, and Yb. Other for ocean-island tholeiitic and alkalic series lavas, but it does appear elements analyzed were Zr, Y, Sr, V, Ti, and Cr. These data are to be diagnostic for basanitic series lavas that have clinopyroxene presented in Table 5. Analytical uncertainty is estimated at 20%-30% grains consistently containing greater than 10% tetrahedral substitu- for Eu and 15% for all other elements. Reflected-light photomicro- tion by Al (Al(z) > 10) (Fig. 2A, D). graphs were used during ion probe analyses to locate spots previously analyzed by electron probe. Thus, each ion probe analysis has a cor- The third diagram, also from Le Bas (1962), is a Fe+Mn, Mg, responding electron probe analysis less than 50 µm away. These Ca+Na ternary (Fig. 3). This is essentially the classic Fs-En-Wo specific locations are denoted as CPX A, CPX B, and so on for each pyroxene ternary, but the addition of Na and Mn better separates the xenolith ("CPX" = clinopyroxene). nonalkaline, alkaline, and peralkaline fields. This diagram is very useful for mafic (normalized Fe+Mn < 20%) groundmass composi- COGENETIC MINERALS tions. Most phenocryst compositions from a given lava type will fall into the respective field, but some will scatter across the nearest Major Element Chemistry boundary (Le Bas, 1962). Despite the scatter resulting from electron microprobe analyses, groundmass clinopyroxene compositions from Olivine the Hawaiian tholeiitic, alkalic, and nephelinitic series show only a The only Leg 144 lavas with unaltered olivine phenocrysts are the slight overlap (Fig. 3A), indicating that this diagram is a useful ankaramitic lavas of Wodejebato Guyot (Hole 874B) and one alkalic diagnostic tool for hotspot lavas.
514 MINERAL COMPOSITIONS OF LAVAS AND XENOLITHS
Table 1. Minerals analyzed by electron and/or ion microprobe.
Range of unit Groundmass Phenocrysts Xenoliths Core, section, Igneous (core, section, Source CPX PLAG CPX OL PLAG CPX OPX OL interval (cm) unit interval [cm]) Limalok Guyot: 144-871C- 144-871C- 35R-1, 10-12 Author Unitl 35R-1,6-58 E E,l 35R-1, 14-16 Author Unit 1 35R-1,6-58 E 35R-1,20-23 A Author Unit 1 35R-1,6-58 E.I 35R-1.20-23B Author Unit 1 35R-1,6-58 E E E E 35R-1,37-39 Author Unit 1 35R-1,6-58 E,I E.I 38R-1,29-32 Author Unit 17 38R-1, 18to38R-4, 17 E,I E E,I 38R-3, 28-30 Author Unit 17 38R-1, 18to38R-4, 17 E E 38R-6, 52-55 Shipboard Unit 19 38R-5, 30 to 38R-6, 135 E 40R-1, 141-143 Shipboard Unit 22B 40R-1, 140 to 40R-2, 53 E E Lo-En Guyot: 144-872B- 144-872B- 5R-1,76-78 Author Unit 6 4R-l,33to5R-2, 6 E E 5R-3, 114-115 Author Unit 9B 5R-3, 68 to 5R-4, 23 E E,I 6R-1,75-80 Shipboard Unit 11 6R-1,74-100 E E 7R-5, 93-95 Author Unit 14B 7R-4, 56 to 7R-6, 43 E E E 7R-7, 23-25 Author Unit 15B 7R-6, 135to7R-7, 83 E E 8R-2, 131-133 Author Unit 16B 8R-2, 18to9R-l,5 E E 9R-2, 92-94 Shipboard Unit 17B 9R-2, 81 to9R-3,90 E 9R-5, 107-109 Author Unit 18B 9R-4, 0 to 9R-6, 60 E,I 9R-6, 14-15 Author Unit 18B 9R-4, 0 to 9R-6, 60 E 9R-6, 41-42 Author Unit 18B 9R-4, 0 to 9R-6, 60 E Wodejebato Guyot 144-873A- 144-873A- 15R-1,23-26 Shipboard Unit 4 15R-1,2-54 E 17R-1,22-28 Shipboard Unit 5 17R-1,0-39 E E 18R-1,30-33 Author Unit 8 18R-1, 0-133 E E,I 18R-1,94-95 Author Unit 8 18R-1, 0-133 E E,I 144-874B- 144-874B- 23R-1,23-25 Author Unitl 22R-4, 0 to 24R-4, 130 E 24R-1, 12-16 Author Unit 1 22R-4, 0 to 24R-4, 130 E E.I E 24R-4, 11-15 Author Unit 1 22R-4, 0 to 24R-4, 130 E E.I E E 24R-4, 102-106 Author Unit 1 22R-4, 0 to 24R-4, 130 E
144-875C- 144-875C- 15M-1, 1-4 Author Unit 1 14M-l,65to 15M-1.77 E E 15M-1, 19-21 Author Unit 1 14M-l,65tol5M-l,77 E 144-876A- 144-876A- 16R-1, 1-3 Author Unit IB 15R-l,56to 16R-1,35 E E E 16R-1,9-11 Author Unit IB 15R-l,56to 16R-1.35 E 17R-1,56-61 Author Unit 3 17R-1,40-96 E E,I 144-877A- 144-877A- 20R-5, 22-24 Author Unitl 20R-4, 98 to 20R-5, 35 E E,I E MIT Guyot: 144-878A- 144-878 A- 78R-2, 44-46 Author Unitl 78R-l,44to78R-2, 88 E E 80R-6, 101-104 Author Unit 8 80R-2, 74 to 80R-6, 120 E 81R-2, 37-40 Author Unit 10 81R-l,6to81R-3, 103 E E E,l 81R-5,22-25 Shipboard Unit 11 81R-3, 103to82R-l,0 E 81R-5,60-62 Author Unit 11 81R-3, 103to82R-l,0 E E 83R-1,6-8 Author Unit 13 83R-2, 0 to 83R-4, 43 E E E.I E 84R-4, 111-113 Author Unit 15 84R-1, 19to85R-l,0 E E 88R-3, 42-44 Author Unit 20 88R-l,0to88R-3, 97 E E 89R-2, 42-43 Author Unit 21 89R-l,0to90R-2, 70 E E,I E 89R-2, 87-89 Author Unit 21 89R-l,0to90R-2, 70 E E 89R-2, 101-103 Author Unit 21 89R-l,0to90R-2, 70 E E 89R-2, 133-137 Author Unit 21 89R-l,0to90R-2,70 E 89R-3, 113-115 Author Unit 21 89R-l,0to90R-2, 70 E 90R-5, 111-116 Author Unit 23 90R-4, 145 to 92R-1, 0 E,I 90R-6, 63-64 Author Unit 23 90R-4, 145to92R-l,0 E E,I E 91R-1, 88-89 Author Unit 23 90R4, 145to92R-l,0 E E,I 91R-3.79-81 Author Unit 23 90R-4, 145to92R-l,0 E 91R-3,94-97 Shipboard Unit 23 90R4, 145to92R-l,0 E E 91R-5,47-54 Author Unit 23 90R4, 145to92R-l,0 E 92R-5, 41-43 Shipboard Unit 27 92R-4, 101 to92R-5, 86 E 93R-2, 111-114 Author Unit 30 93R-2, 63to94R-l,58 E E 95R-6, 33-34 Author Unit 32 95R-5, 83 to 95R-6, 58 E 97R-1,46-50 Author Unit 34 97R-1, 19to98R-l,0 E 97R-2, 113-115 Author Unit 34 97R-1, 19to98R-l,0 E 97R-2, 128-130 Author Unit 34 97R-1, 19to98R-l,0 E E 97R-3, 5-10 Author Unit 34 97R-1, 19to98R-l,0 E Takuyo-Daisan Guyot: 144-879A- 144-879A- 22R-1, 133-135 Author Unit IB 21R-l,65to24R-l, 106 E,I 24R-1, 85-87 Author Unit IB 21R-l,65to24R-l, 106 E E E E
Notes: CPX = clinopyroxene, PLAG = plagioclase, OL = olivine, and OPX - orthopyroxene. "E" implies electron probe analyses of the phase were done, and "I" implies ion probe analyses of the phase were done.
515 J.J. DIEU
Table 2. Representative olivine analyses.
Hole: 874B 874B 874B 874B 874B 874B 874A 874A 874A 874A 874A 874A 874A Core, section: 24R-1 24R-1 '24R-4 24R-4 24R-4 24R-4 84R-4 84R-4 90R-6 90R-6 91R-5 91R-5 91R-5 Interval (cm): 12-16 12-16 12-16 11-15 11-15 11-15 111-113 111-113 63-64 63-64 47-54 47-54 47-54 Description: Core of small Core of Small Small Large Small Small Small Small Small Dunite Dunite Dunite phenocryst phenocryst phenocryst phenocryst phenocryst phenocryst phenocryst phenocryst phenocryst phenocryst xenolith xenolith xenolith Analysis: OL5 OL6 OL2 OL7 OL 18 OL23 OL2 OL6 OL5 OL6 OL2 OL3 OL7 Analysis no.: 1 2 3 5 8 10 11 12 13 14 15 16 17 Element (ppm) SiO2 39.60 39.08 38.64 38.71 38.84 40.15 39.26 39.24 39.26 37.73 39.39 38.76 36.97 TiO2 0.05 0.02 0.00 0.00 0.02 0.00 0.02 0.03 0.01 0.02 0.03 0.02 0.04 A12O3 0.04 0.03 0.04 0.03 0.05 0.04 0.07 0.05 0.05 0.03 0.01 0.00 0.03 Cr2O3 0.04 0.04 0.02 0.01 0.04 0.03 0.03 0.04 0.00 0.01 0.00 0.00 0.02 FeO 15.17 14.82 18.00 18.52 14.69 13.76 14.70 14.83 15.79 26.11 14.05 19.44 26.82 MnO 0.19 0.19 0.24 0.26 0.18 0.20 0.22 0.22 0.22 0.36 0.19 0.36 0.39 MgO 44.77 44.80 42.94 42.56 46.15 46.50 45.37 45.26 44.89 37.53 46.24 42.33 36.81 NiO 0.21 0.18 0.18 0.18 0.15 0.19 0.23 0.22 0.32 0.07 0.40 0.27 0.09 CaO 0.24 0.30 0.31 0.31 0.22 0.34 0.27 0.28 0.15 0.20 0.10 0.08 0.18 Total 100.31 99.46 100.37 100.58 100.34 101.21 100.17 100.17 100.69 102.06 100.41 101.26 101.35 Per four oxygens: Si 0.994 0.989 0.984 0.986 0.974 0.992 0.986 0.986 0.986 0.981 0.984 0.984 0.973 Ti 0.000 0.000 0.000 0.000 0.000 0.000 0.000 0.000 0.000 0.000 0.000 0.000 0.000 Al 0.001 0.000 0.001 0.000 0.001 0.001 0.002 0.001 0.000 0.000 0.000 0.000 0.000 Cr , 0.000 0.000 0.000 0.000 0.000 0.000 0.000 0.000 0.000 0.000 0.000 0.000 0.000 Fe+2 0.318 0.314 0.383 0.394 0.308 0.284 0.309 0.312 0.332 0.567 0.294 0.413 0.590 Mn 0.004 0.004 0.005 0.006 0.004 0.004 0.005 0.005 0.005 0.008 0.004 0.008 0.009 Mg 1.675 1.690 1.630 1.615 1.726 1.712 1.698 1.695 1.680 1.454 1.722 1.602 1.444 Ni 0.004 0.004 0.004 0.004 0.003 0.004 0.005 0.004 0.006 0.001 0.008 0.006 0.002 Ca 0.006 0.008 0.008 0.008 0.006 0.009 0.007 0.008 0.004 0.006 0.003 0.002 0.005 Total 3.002 3.009 3.015 3.013 3.022 3.006 3.012 3.011 3.013 3.017 3.015 3.015 3.023 Fo 84.0 84.3 81.0 80.4 84.8 85.8 84.6 84.5 83.5 71.9 85.4 79.5 71.0
Note: Analysis numbers refer to complete data table on CD-ROM (back pocket).
Only a part of this table is reproduced here. The entire table appears on the CD-ROM (back-pocket).
Table 3. Representative clinopyroxene and orthopyroxene analyses.
Limalok Guyot
Hole: 871C 871C 871C 871C 871C 871C 871C 871C Core, section: 35R-1 35R-1 35R-1 35R-1 35R-1 38R-1 38R-1 38R-3 Interval (cm): 10-12 14-16 20-23 23-25 37-39 29-32 29-32 28-30 (Xenolith A) (Xenolith B) Description: Alkalic Phenocryst Mantle Mantle Alkalic Core of Rim of Phenocryst cumulate residue residue cumulate 1.0-cm 1,0-cm xenolith xenolith xenolith xenolith xenocryst; xenocryst; grain grain grain grain CPX 4, 6, 7, CPX 4, 5, 7, 8, A, B 8, A, B Analysis: CPX 1 CPX 14 CPXB OPX7 CPX 9 CPX 5 CPX 6 CPX 10 Analysis no.: 1 14 21 28 31 39 40 51
Element (ppm): SiO2 51.74 44.89 51.93 55.45 49.17 43.14 46.23 47.15 TiO2 1.09 4.10 0.14 0.01 1.28 3.58 2.22 2.26 A12O3 4.60 8.79 6.64 3.78 8.49 11.70 9.13 6.42 Cr2O3 0.31 0.03 1.30 0.46 0.02 0.00 0.41 0.05 FeO 5.68 7.12 2.62 5.99 5.98 7.65 6.08 7.01 MnO 0.09 0.03 0.09 0.10 0.06 0.05 0.19 0.05 MgO 18.56 13.00 16.21 32.99 13.82 11.36 13.33 13.74 NiO 0.02 0.05 0.04 0.11 0.00 0.01 0.08 0.01 CaO 18.23 21.94 19.77 1.18 20.51 21.97 22.14 22.23 Na2O 0.60 0.33 1.17 0.18 1.10 0.81 0.59 0.37 Total 100.91 100.28 99.91 100.25 100.43 100.27 100.40 99.28 Per six oxygens: Si 1.867 1.676 1.875 1.913 1.799 1.617 1.713 1.772 Ti 0.030 0.115 0.004 0.000 0.035 0.101 0.062 0.064 Al 0.196 0.387 0.283 0.154 0.366 0.517 0.399 0.284 Cr 0.009 0.000 0.037 0.013 0.000 0.000 0.012 0.001 Fe+2 0.171 0.222 0.079 0.173 0.183 0.240 0.188 0.220 Mn 0.002 0.000 0.003 0.003 0.002 0.002 0.006 0.002 Mg 0.999 0.723 0.872 1.696 0.754 0.635 0.736 0.770 Ni 0.000 0.002 0.001 0.003 0.000 0.000 0.002 0.000 Ca 0.705 0.877 0.765 0.044 0.804 0.882 0.879 0.895 Na 0.042 0.024 0.082 0.012 0.078 0.059 0.042 0.027 Total 4.022 4.028 4.000 4.010 4.021 4.053 4.039 4.035 En 53.26 39.68 50.83 88.66 43.31 36.14 40.82 40.85 Fs 9.14 12.19 4.61 9.04 10.51 13.66 10.43 11.67 Wo 37.60 48.13 44.56 2.30 46.18 50.20 48.75 47.48
Note: Analysis numbers refer to clinopyroxene data tables, by guyot, on the CD-ROM (back pocket).
Only a part of this table is reproduced here. The entire table appears on the CD-ROM (back-pocket).
516 MINERAL COMPOSITIONS OF LAVAS AND XENOLITHS
o t>.u- α Guyot phenocrysts D Lo-En o Lo-En Guyot groundmass α Limalok Guyot D 4.0- phenocrysts G / 4.0- o Limalok Guyot / groundmass O o Ed / / 3.0- o / 3.0- / ° o CM / / CM DG] O O α/ O I- S 2.0- / / 2.0- o / α ^-/ ×^? — <> o /// -^ 1.0- — —'-*- 1.0- — —
0.0- 0.0- T r~ —I 0.90 0.85 0.80 0.75 0.70 0.90 0.85 0.80 0.75 0.70 Mg/(Mg+Fe) Mg/(Mg+Fe)
5.0 5.0 Wodejebato Guyot D Site 873 O Site 874 4.0- 4.0 -i MIT Guyot alkalic phenocrysts O Site 875 MIT Guyot alkalic groundmass A Site 876 ffl Site 877 3.0- 3.0-
CM CN o o 2.0- 2.0-
1.0- 1.0-
MIT Guyot basanitic phenocrysts MIT Guyot basanitic groundmass 0.0 0.0 0.90 0.85 0.80 0.75 0.70 0.90 0.85 0.80 0.75 0.70 Mg/(Mg+Fe) Mg/(Mg+Fe) £ 5.0
4.0- /
3.0- /
CM / / O 2.0-
o ^ 1.0-
α Takuyo-Daisan Guyot phenocrysts o Takuyo-Daisan Guyot groundmass 0.0- 0.90 0.85 0.80 0.75 0.70 Mg/(Mg+Fe)
2+ Figure 1. TiO2 vs. Mg/(Mg+Fe) of clinopyroxene phenocrysts. Mg and Fe calculated on six oxygens. All Fe in the clinopyroxene analyses are assumed to be Fe . Fields on each of Figures 1A-E are as follows: solid line = Hawaiian tholeiitic phenocrysts, and dashed line = Hawaiian alkalic phenocrysts. Hawaiian fields are drawn from data in Fodor et al. (1975), Clague et al. (1980), Garcia et al. (1982, 1986), Ho and Garcia (1988), Frey et al. (1990), and Chen et al. (1990).
517 J.J. DIEU
B
15.0- 0 Peralkaline 15.0- Peralkalihe
£ 10.0H r 10.0-
5.0- 5.0-
Limalok Guyot phenocrysts Lo-En Guyot phenocrysts Limalok Guyot groundmass Lo-En Guyot groundmass 0.0 0.0 0.0 1.0 2.0 3.0 4.0 5.0 0.0 1.0 2.0 3.0 4.0 5.0
Ti02 (%) TiO2 (%)
Peralkaline 15.0- \ 15.0- Peralkaline
£ 10.0H / =: 10.0-
Wodejebato Guyot 5.0- π Site 873 5.0- O Site 874 O Site 875 MIT Guyot alkalic phenocrysts MIT Guyot alkalic groundmass Δ Site 876 MIT Guyot basanite phenocrysts ffl Site 877 MIT Guyot basanite groundmass 0.0 0.0 0.0 1.0 2.0 3.0 4.0 5.0 0.0 1.0 2.0 3.0 4.0 5.0 Ti02 (%) Ti02 (%)
20.0
15.0-
^ 10.0- <
5.0-
o Takuyo-Daisan Guyot phenocrysts o Takuyo-Daisan Guyot groundmass 0.0 0.0 1.0 2.0 3.0 4.0 5.0 Ti02
IV Figure 2. Al(z) vs. TiO2 of clinopyroxene phenocrysts. Al(z) = A1 × 100/2. Al(z) and field boundaries from Le Bas (1962) for groundmass compositions. Fields on each of Figures 2A-E as on Figure 1. Field boundaries do not represent absolute delineations for phenocrysts. Diagonal lines without arrows (on each of Fig. 2A-E) are divisions between nonalkaline/alkaline and alkaline/peralkaline fields as established by Le Bas (1962). Lines with arrows (on Fig. 2B only) are toward increasing differentiation for nonalkaline and alkaline/peralkaline lavas in the respective divisions.
518 MINERAL COMPOSITIONS OF LAVAS AND XENOLITHS
Fe + Mn 10 15 20
55 50 45 55 50 45 55 50 45
MIT Guyot ( Hawaiian T Limalok Guyot α tholeiitic series basanitic series • Lo-En Guyot Hawaiian MIT Guyot alkalic series Wodejebato alkalic series Guyot ^ Hawaiian p. Takuyo-Daisan nephelinitic series Guyot
Figure 3. Fe+Mn, Mg, Ca+Na ternary diagram. Values are calculated from structural formulas based on six oxygens. Dashed lines separate nonalkaline/alkaline (lower line) and alkaline/peralkaline (upper line) boundaries after Le Bas (1962). All data are from groundmass analyses. Hawaiian data from sources listed in Figure 1 caption.
Limalok Guyot Lavas (1962) for nonalkaline lavas. However, other groundmass composi-
tions plot at higher Al(z) and TiO2 wt% than do the phenocryst Petrographic and major element analyses of the lavas recovered in compositions, suggesting a differentiation trend like that observed by Hole 871C at Limalok Guyot unequivocally place these lavas in the Le Bas (1962) for alkaline and peralkaline magmas. basanitic series (Christie et al., this volume). This strongly alkalic Petrographic observations, trace element geochemistry (Christie et nature is reflected in the clinopyroxene chemistry. The TiO2 contents al., this volume), and the tight phenocryst correlation on Figure IB all of phenocrysts range from 1.6 to 4.6 wt% and are negatively correlated suggest that the Lo-En Guyot lavas belong to a single series (i.e., they with Mg#, which ranges from 0.74 to 0.82 (Fig. 1A; Table 3). The were derived from essentially identical parental liquids). Thus, results compositions of groundmass grains vary more widely, but most lie on Figure 2B are difficult to interpret. However, the cluster of Lo-En within the same range as the phenocryst compositions. Most of these Guyot groundmass compositions on the Fe+Mn, Mg, Ca+Na ternary phenocryst and groundmass compositions have greater TiO2 for a (Fig. 3B) straddle the nonalkaline/alkaline boundary, suggesting an given Mg# than do Hawaiian alkalic series compositions (Fig. 1A). obvious interpretation: the Lo-En Guyot lavas are transitional between Despite the mafic nature of the lavas, most of the clinopyroxene com- tholeiitic and alkalic. These lavas plot in the alkalic region of the SiO2 positions plot in the peralkaline field on the Al(z) vs. TiO diagram of 2 vs. Na2O+K2O diagram (Christie et al., this volume); however, they lie
Le Bas (1962) and many lie at higher TiO2 and Al(z) than the Hawaiian very close to the tholeiitic/alkalic boundary and, as these lavas are alkalic series compositions (Fig. 2A). This suggests that the Limalok somewhat altered, their Na and K abundances must be viewed as Guyot lavas are more alkalic than the Hawaiian alkalic series lavas. suspect. Thus, their position on the silica-alkalis diagram cannot be This suggestion is supported by the range of Limalok Guyot viewed as strongly diagnostic, especially as unaltered clinopyroxene groundmass clinopyroxene compositions plotted on Figure 3B. As compositions support an interpretation of their being transitional. might be expected of basanitic series lavas, the groundmass clinopy- roxene compositions of Limalok Guyot plot between (and overlap Wodejebato Guyot Lavas into each of) the fields of Hawaiian alkalic and nephelenitic series groundmass compositions across the alkaline/peralkaline boundary Clinopyroxene is present in the altered, alkalic series lavas of each of the Fe+Mn, Mg, Ca+Na ternary (Fig. 3A, B). of the five sites (Sites 873-877) drilled on Wodejebato Guyot. Clino- pyroxene compositions from the ankaramitic lavas of Site 874 are Lo-En Guyot Lavas clearly more mafic, but phenocryst data from all sites lie along a
well-defined fractionation trend in a plot of TiO2 vs. Mg# (Fig. 1C), Hole 872B lavas have been interpreted as belonging to a moder- suggesting that the same lava series is present at all sites (see also ately differentiated alkalic series (Christie et al., this volume). Con- Christie et al., this volume). The only exceptions are two low-Ti sistent with this interpretation, clinopyroxene compositions display clinopyroxene phenocrysts from Sites 874 and 876. The Site 874
moderate TiO2 contents (0.6%-1.3%) and a tight, negative correla- analysis is of a rim of a large, euhedral, zoned phenocryst; the Site 876
tion, usually interpreted as being fractionation controlled, on the TiO2 analysis is of a small, euhedral phenocryst. Neither of these grains is vs. Mg# diagram of Figure IB. obviously a xenocryst. Therefore, it is difficult to explain why these two analyses plot below the overall Wodejebato Guyot trend on In a plot of Al(z) vs. TiO2, Lo-En Guyot clinopyroxene composi-
tions straddle the nonalkaline/alkaline boundary of Le Bas (1962) (cf. Figure 1C. On the Al(z) vs. TiO2 diagram (Fig. 2C), the Wodejebato Fig. 2A, B) and generally lie within the field of Hawaiian alkalic Guyot clinopyroxene phenocryst compositions cover a long, narrow series lavas. However, several of the groundmass compositions lie at swath across the nonalkaline/alkaline boundary and do not com- pletely fall into either of the Hawaiian fields. Note that this long, nar- lower Al(z) and TiO2 wt% (Fig. 2B) than do the phenocryst composi-
tions, defining a differentiation trend like that observed by Le Bas row swath is clearly a differentiation trend (from the TiO2 vs. Mg#
519 J.J. DIEU relationship on Fig. 1C) paralleling the direction of alkaline and guyot straddle the alkaline/peralkaline boundary on Figure 3. These peralkaline magma differentiation shown on Figure 2B. Phenocrysts compositions are clearly quite distinct from those of other Leg 144 with compositions that plot below the Hawaiian alkalic series field lavas and may be diagnostic characteristics of clinopyroxene compo- crystallized from the mafic ankaramitic lavas of Hole 874B. As these sitions from basanitic series lavas. lavas are more mafic than any of the Hawaiian alkalic series lavas and Alkalic series lavas were recovered from both MIT and Wodejebato as this trend parallels the alkaline/peralkaline differentiation trend of guyots. Some lavas from Wodejebato Guyot have unusually mafic Le Bas (1962), the interpretation that all Wodejebato Guyot lavas are compositions, even when compared to a large set of Hawaiian lavas. alkalic is consistent with these compositions. The Wodejebato lavas appear to represent a large range of differentia- The Wodejebato Guyot groundmass clinopyroxene compositions tion of a single parental magma type. Conversely, the MIT Guyot alkalic cover a range similar to those from Lo-En Guyot on Figure 3B, series lavas are all moderately fractionated and appear to have been de- although only one Wodejebato Guyot analysis falls significantly be- rived from several distinct parental magma types (or possibly several low the nonalkaline/alkaline boundary. Therefore, I conclude that the distinct magma chamber depths; these two effects cannot be separated). Wodejebato Guyot lavas are mildly alkalic because most of the clino- If only the phenocryst compositions are examined, Lo-En Guyot pyroxene analyses plot in the alkaline field on Figure 3B and display clinopyroxene grains form the tightest cluster of values on a plot of an alkaline differentiation trend on Figure 2C. Those compositions TiO2 vs. Mg# (Fig. IB), indicating that the sequence of erupted lavas that lie below the Hawaiian alkalic series field on Figure 2C are sampled by drilling represents a narrow range of differentiation of a simply more mafic than any Hawaiian compositions reported in the single parental magma type. This parental magma type was clearly literature (Fig. 1C) and do not indicate that the Wodejebato Guyot distinct from the types represented by the Wodejebato and MIT guyot lavas are tholeiitic or transitional. lavas. The clinopyroxene compositions indicate that the Lo-En Guyot lavas are transitional between alkalic and tholeiitic. Specifically, the MIT Guyot Lavas phenocryst compositions have low TiO2 (Figs. IB, 2B), and the ground- mass compositions plot across the nonalkaline/alkaline boundary of The lavas recovered from Site 878 (MIT Guyot) have been inter- Figure 3B. preted as subaerial basanitic and moderately fractionated alkalic se- The single lava from Takuyo-Daisan Guyot contains clinopyrox- ries flows (Premoli Silva, Haggerty, Rack, et al., 1993; Christie et al., ene phenocrysts and groundmass grains with compositions that are this volume). As with the Limalok Guyot lavas, the MIT Guyot lavas similar to those of Lo-En Guyot lavas. This lava may be transitional identified as basanitic have clinopyroxene grains with > 2 wt% TiO2 in nature, like the Lo-En Guyot lavas, but petrographic evidence (Figs. ID, 2D), and most compositions have > 10% tetrahedral alu- (Christie et al., this volume) and plagioclase compositions (discussed minum substitution and thus plot in the peralkaline field of Figure 2D. below) cannot preclude a tholeiitic nature. Also, the MIT Guyot analyses plot across the alkaline/peralkaline boundary of Figure 3, as do the Limalok Guyot basanitic groundmass Feldspar clinopyroxene compositions. The MIT Guyot lavas identified as belonging to the alkalic series Feldspar compositions are diagnostic of the degrees of differentia- are moderately fractionated. The clinopyroxene Mg#s range from tion and silica-saturation of a lava. Detailed criteria were established for Hawaiian lavas by Keil et al. (1972); these are summarized here. 0.82 to 0.73, and there is little correlation with TiO2 (Fig. ID). Pheno- cryst compositions plot across both the nonalkaline/alkaline and al- Simply stated, tholeiitic lavas do not contain feldspar phenocrysts, kaline/peralkaline boundaries on Figure 2D, but most fall in the microphenocrysts, or groundmass laths with Or > 1.1, and no tholei- alkaline field. The groundmass compositions straddle the nonalka- itic feldspar compositions plot in the two-feldspar region or in the line/alkaline boundary on Figure 3C, but no analyses fall very far into anorthoclase or sanidine regions of the An-Ab-Or ternary (Fig. 4A). the nonalkaline field. Collectively, these diagrams suggest that the Alkalic and "nephelinic" (equivalent to basanitic; Keil et al., 1972) nonbasanitic MIT Guyot lavas are correctly identified by Christie et lavas contain at least some groundmass laths and interstitial material al. (this volume) as essentially alkalic series lavas. However, some (crystalline or glassy with feldspar compositions) of anorthoclase or may be transitional in nature, and, probably, they were not derived sanidine compositions. Some alkalic lavas contain feldspar grains from a single parental magma type. As there are long erosional epi- with compositions in the two-feldspar region (Fig. 4A). sodes recorded in the igneous basement of Hole 87 8 A, the latter is not As presented in Table 4 (and on the CD-ROM), definitions of a surprising result. phenocryst and microphenocryst are identical to those of Keil et al. (1972), but definitions of groundmass and interstitial are not equiva- Takuyo-Daisan Guyot Lavas lent. Specifically, I do not have an interstitial category, although "poikilitic in groundmass" is interstitial material as Keil et al. (1972) A peperite formed by eruption of clinopyroxene-plagioclase-phyric defined it. However, most analyses of groundmass grains (Table 4 and basalt into unconsolidated volcaniclastic sediment was the only igne- CD-ROM) are of laths because interstitial feldspar grains and glassy ous unit recovered at Site 879. Clinopyroxene Mg#s range from 0.77 material in Leg 144 lavas have been altered to clay and chlorite to 0.81, with low Al and Ti abundances (Figs. IE, 2E). The single minerals. Therefore, the criteria of Keil et al. (1972) may be used with groundmass composition falls well below the nonalkaline/alkaline little attention to differences in definitions of grain types. boundary on Figure 3B and even below the Lo-En Guyot transitional series field. The clinopyroxene compositions suggest that Hole 879A Limalok Guyot Lavas lava is a moderately fractionated transitional or tholeiitic basalt. The basanitic lavas from Site 871 do not contain plagioclase Comparisons Among Clinopyroxene Results phenocrysts. All compositions plotted on Figure 4A are of microlites of Individual Guyots and poikilitic groundmass grains. Values range from An619 to An25 2 and Or increases smoothly with decreasing An. That these lavas are The basanitic lavas of Limalok and MIT guyots crystallized clino- alkalic is supported by an analysis of a microphenocryst with Or = pyroxene grains of similar and rather remarkable chemistry (Table 3). 8.69 (Table 4, analysis number 12). These compositions are distinct from alkalic series lavas, having Despite thorough exploration, no feldspar of sanidine or anortho- higher TiO2 (usually > 2%) and higher Al (typically with > 10% tetra- clase composition was found during analysis. Either potassic feldspar hedral aluminum substitution) (Figs. 1-3). Furthermore, the ground- did not crystallize in the groundmass or, as seems likely from petro- mass clinopyroxene compositions of the basanitic lavas from each graphic observations, it completely altered to clay minerals. Whole-
520 MINERAL COMPOSITIONS OF LAVAS AND XENOLITHS
B
• Lo-En Guyot phenocrysts Limalok Guyot groundmass » Lo-En Guyot groundmass
Wodejebato Guyot • Site 873 phenocrysts MIT Guyot phenocrysts T Site 873 groundmass MIT Guyot groundmass ^ Site 874 groundmass A Site 875 groundmass A Site 876 groundmass • Site" 877 phenocrysts ° Site 877 groundmass
a Takuyo-Daisan phenocrysts r Takuyo-Daisan groundmass
Figure 4. Orthoclase-albite-anorthite feldspar ternary diagrams. Fields on Figure 4A are from Deer et al. (1966). rock K2O contents are unusually low for basanitic lavas and low with a smooth trend toward increasing Or with decreasing An. Ground- relative to the abundances of trace elements of similar incompatibility, mass grains, primarily microlites, range from An613 to Anu 2 and suggesting significant losses by alteration (Christie et al., this volume). several of the analyses plot outside of the fields of equilibrium feld- spar compositions (Fig. 4B) in the two-feldspar field. Both the high- Lo-En Guyot Lavas Or phenocrysts and the microlites with two-feldspar compositions The moderately differentiated transitional lavas recovered from attest to the alkalic nature of these lavas. However, by their very Lo-En Guyot contain plagioclase and alkali feldspar phenocrysts with nature, transitional lavas have some characteristics in common with compositions that vary from An82 5 to An9 5 and from Or0 3 to Or48 3, tholeiitic lavas (e.g., Hawaiian transitional lavas often have mineral
521 J.J. DIEU
Table 4. Representative plagioclase analyses.
Hole: 871C 871C 871C 871C 872B 872B 872B 872B 872B 872b Core, section: 35R-1 38R-I 38R-6 40R-1 5R-1 5R-1 5R-1 6R-1 6R-1 7R-5 Interval (cm): 14-16 29-32 52-55 141-143 76-78 76-78 76-78 75-80 75-80 93-95 Description: Poikilitic in Poikilitic in Poikilitic in Micro- Rim of large Core of large Ground- Ground- Micro- Ground- groundmass groundmass groundmass phenocryst phenocryst; phenocryst; mass mass phenocryst mass PL 1,3 PL 1,2 (50 µm) Analysis: PL 16 PL 11 PL 2 PL 3 PL 2 PL 3 PL 4 PL 6 PL 12 PL 11 Analysis no.: 2 5 8 12 14 15 16 19 24 29
Element (ppm):
SiO2 52.36 62.00 52.54 59.38 47.96 47.95 54.79 51.82 62.69 53.06 TiO2 0.31 0.11 0.07 0.09 0 0.09 0.11 0.18 0.78 0.18
A12O3 28.17 22.36 30.40 24.31 33.31 33.82 27.88 28.33 20.18 29.23 FeO ' 1.27 0.82 0.37 0.41 0.42 0.44 1.39 1.96 1.78 0.94 MnO 0.01 0.01 0.01 0.00 0.01 0.03 0.01 0.01 0.05 0.01 MgO 1.89 1.01 0.03 0.46 0.14 0.08 0.51 0.89 0.19 0.10 CaO 10.83 4.08 12.68 5.96 17.2 17.31 10.75 12.33 2.04 12.54 Na,O 4.51 7.36 4.15 7.14 2.13 1.99 5.03 4.13 6.48 4.57
K2O 0.30 2.24 0.25 1.51 0.09 0.06 0.7 0.27 6.48 0.32 Total 99.65 99.99 100.50 99.26 101.26 101.77 101.17 99.92 100.67 100.95 Per 32 oxygens: Si 9.572 11.094 9.496 10.725 8.724 8.676 9.861 9.511 11.357 9.589 Al 6.069 4.716 6.475 5.175 7.141 7.212 5.914 6.128 4.309 6.225 Fe 0.194 0.123 0.056 0.062 0.064 0.067 0.209 0.301 0.270 0.142 Mπ 0.002 0.002 0.002 0.000 0.002 0.005 0.002 0.002 0.008 0.002 Ms 0.515 0.269 0.008 0.124 0.038 0.022 0.137 0.244 0.051 0.027 Ca 2.121 0.782 2.455 1.153 3.352 3.356 2.073 2.425 0.396 2.428 Na 1.599 2.553 1.454 2.500 0.751 0.698 1.755 1.470 2.276 1.601 K 0.070 0.511 0.058 0.348 0.021 0.014 0.161 0.063 1.498 0.074 Total 20.185 20.065 20.013 20.099 20.092 20.061 20.126 20.167 20.270 20.112 An 55.97 20.33 61.89 28.82 81.28 82.50 51.97 61.27 9.50 59.18 Ah 42.18 66.38 36.66 62.48 18.21 17.16 44.00 37.14 54.59 39.03 Or 1.85 13.29 1.45 8.69 0.51 0.34 4.03 1.60 35.92 1.80
Note: Analysis numbers refer to complete data table on CD-ROM (back pocket).
Only a part of this table is reproduced here. The entire table appears on the CD-ROM (back-pocket).
assemblages identical to Hawaiian tholeiitic lavas (Macdonald and such huge abundances suggests that plagioclase was not yet on the Katsura, 1964) and some characteristics in common with alkalic liquidus, a crystallization sequence characteristic of alkalic liquids. lavas. It is not unusual for feldspar chemistry to reflect the alkalic characteristics of transitional lavas (e.g., Keil et al., 1972). MIT Guyot Lavas
Wodejebato Guyot Lavas The moderately differentiated, alkalic series lavas recovered from MIT Guyot contain plagioclase phenocrysts with compositions rang-
Plagioclase compositions from the alkalic series lavas of Sites ing from An818 to An58 9 (Fig. 4D). This apparently small range for 873-877 (Wodejebato Guyot) show a broad variability in degree of such a petrographically diverse suite of lavas (see Premoli Suva, differentiation, similar to that shown by the clinopyroxene composi- Haggerty, Rack, et al., 1993) probably indicates that the lavas under- tions. Site 874 lavas do not contain plagioclase phenocrysts, but the went similar degrees of differentiation. Of the lavas identified as
microlite compositions are quite anorthitic (An78 3 to An64 5). These alkalic series, analyses of groundmass laths with Or > 1.1 exist for indicate crystallization from a relatively primitive liquid (Fig. 4C). igneous Units 1,13, and 23, the lava of igneous Unit 27 crystallized The Site 877 lava is similarly undifferentiated, with phenocryst com- groundmass grains of anorthoclase composition, and the lava of igne-
positions ranging from An87 7 to An68 8 and microlite compositions ous Unit 34 crystallized microphenocrysts with Or > 1.1 (see Table 4
ranging from An67 7 to An63 3. Site 873 phenocrysts range from An68A and CD-ROM). From the criteria of Keil et al. (1972), it is clear that
to An60 0 and microlites range from An56 9 to An35 2• This lava is more those MIT Guyot units identified as alkalic series lavas did crystallize fractionated than lavas from Sites 874 and 877. The Site 875 compo- feldspar of distinctly alkalic character.
sitions, from An59 4 to An40 3, indicate that the liquid is moderately Those MIT Guyot lavas identified as basanitic crystallized ground-
fractionated, similar to the Site 873 lava. Site 876 compositions cover mass feldspar grains with compositions that range from An67 to An9 6
a broader range (An679 to An218) and include two analyses that plot and from Orj 0 to Or45 4. These basanitic lavas contained few plagio- in the two-feldspar field. Note that only groundmass grains were clase phenocrysts, but all analyses have compositions of Or > 1.1 and analyzed from the Sites 875 and 876 lavas. one analysis plots in the two-feldspar field of the An-Ab-Or ternary
Feldspar compositions that demonstrate the alkalic nature of the at Or35 3. These lavas are clearly alkalic in nature. Wodejebato Guyot lavas were analyzed for lavas from four of the five holes. Hole 873A lavas have phenocrysts and microphenocrysts with Takuyo-Daisan Guyot Lavas Or > 1.1. Holes 875C and 876A lavas have groundmass laths with Or > 1.1. The Hole 877A lava has similar groundmass and phenocryst The clinopyroxene-plagioclase-phyric basalt recovered from rim compositions; and, although none of the analyses has Or > 1.1, Takuyo-Daisan Guyot contains plagioclase phenocrysts with compo-
similar phenocryst rim and groundmass compositions are in them- sitions ranging from An80 7 to An53 4 (Fig. 4E). The only two analyzed selves evidence of crystallization from an alkalic liquid (Keil et al., groundmass grains are within the range of the phenocryst composi-
1972). No analyses for the Hole 874B ankaramitic lavas demonstrate tions, An56 3 and An54 8. One analysis of a phenocryst rim is as potassic
an alkalic nature for these lavas, but the groundmass material is so (Or24) as the groundmass laths (Or23_38), suggesting (on meager altered that few analyses were done. Furthermore, the very lack of evidence) that this lava cannot be tholeiitic. The hypothesis that this plagioclase phenocrysts when clinopyroxene phenocrysts occur in lava is transitional is not precluded by the feldspar data. MINERAL COMPOSITIONS OF LAVAS AND XENOLITHS
Summary of Leg 144 Feldspar Results elements and display convex-upward patterns with slight, positive Ti anomalies (Fig. 5B, D, E). This similarity is rather remarkable be- Feldspar analyses of the Leg 144 lavas demonstrate (merely sug- cause (1) presumably the MIT Guyot lavas were derived from smaller gest in the case of Takuyo-Daisan Guyot) that all of the lavas are degrees of partial melting than the others; (2) isotopic data indicate alkalic or transitional in nature. The occurrences of phenocrysts, that lavas from each of these guyots were derived from distinct mantle microphenocrysts, and groundmass laths with Or > 1.1, of ground- sources; and (3) there is no reason to suppose that the depths of mass grains with anorthoclase, sanidine, or two-feldspar composi- crystallization were equivalent, especially as these volcanoes are of tions, and the phenocryst rims and groundmass grains with similar Or quite different sizes (Premoli Silva, Haggerty, Rack, et al., 1993). components support this conclusion. 3. The alkalic series clinopyroxene phenocrysts from Wodejebato Guyot display a wide range of trace element abundances. Phenocrysts Trace Element Chemistry in Sample 144-874B-24R-1, 12-16 cm, have as low or lower abun- dances of trace elements than do phenocrysts of the transitional lavas Ion microprobe data were collected on 4 microphenocrysts and 23 of Lo-En Guyot (cf. Fig. 5B, C; see Fig. 6). This may be explained, at phenocrysts of clinopyroxene representing all Leg 144 sites except least in part, by differences in degrees of differentiation as the Lo-En Sites 875 and 880. These data are presented in Table 5 and Figures 5-7. Guyot lavas are moderately fractionated and the Hole 874B lavas are relatively primitive (cf. Fig. IB, C), by differences in source enrich- Phenocrysts ment, or by differences in pressure of crystallization. However, no significant differences between the Lo-En Guyot lava source and the The locations of the 23 clinopyroxene phenocryst analyses were Wodejebato Guyot lava source are suggested by the whole-rock and carefully selected on the basis of major element data. Evaluation of isotopic data (Christie et al. and Koppers et al., both in this volume). the phenocryst data is difficult because many factors control the In fact, Lo-En Guyot lavas have, on average, slightly lower 87Sr/86Sr abundances of trace elements in clinopyroxene. First, both the chem- and 143Nd/144Nd ratios and lower Pb isotopic ratios (Koppers et al., istry of the mantle source and the degree of partial melt affect the trace this volume). This suggests that the Lo-En Guyot mantle source had element abundances in magma and, obviously, the abundance of an lower abundances of incompatible trace elements than did the source element in a magma has some affect on the abundance of that element of the Wodejebato Guyot lavas. The reasons for the low trace element in clinopyroxene crystallized from the magma. Second, as a magma abundances in clinopyroxene phenocrysts of Sample 144-874B-24R- differentiates, the abundances of incompatible trace elements in- 1, 12-16 cm, cannot be clearly established. crease and this is reflected in the clinopyroxene compositions. Third, both major element chemistry of a magma and the depth of crystal- However, the low trace element abundances in clinopyroxene phe- lization of a phenocryst control the clinopyroxene/liquid partition nocrysts of Sample 144-874B-24R-1,12-16 cm, are unique among the coefficients of incompatible elements. (Specifically, this has been Wodejebato Guyot samples (Fig. 6). Analyses of other clinopyroxene demonstrated for rare earth elements [REE] and Y; we may reason- phenocrysts show very high abundances of trace elements in compari- ably assume that other incompatible elements are similarly affected son with phenocrysts of other Leg 144 transitional and alkalic series [Hack et al., in press].) These various effects are difficult and some- lavas (Figs. 5, 6). Samples 144-874B-24R-4, 11-15 cm, and 144- times impossible to separate. Consequently, the following discussion 877A-20R-5, 22-24 cm, are relatively primitive (but slightly more is a very simplified interpretation of these data. differentiated than the lava sampled in Sample 144-874B-24R-1, 12- 16 cm). Other Wodejebato lavas are only moderately fractionated. The data plotted by guyot on chondrite-normalized spider dia- Therefore, fractionation cannot explain why most of the Wodejebato grams (Fig. 5) show several interesting features. phenocrysts have higher abundances of trace elements than do pheno- 1. Phenocrysts from basanitic lavas, as analyzed on thin section crysts from the Lo-En, MIT, and Takuyo-Daisan guyot lavas. With the Samples 144-871C-35R-1,37-39 cm, 144-871C-38R-1,29-32 cm, and exception of Sample 144-874B-24R-1, 12-16 cm, the Wodejebato 144-878 A-89R-2,42-43 cm, contain greater abundances of incompat- Guyot clinopyroxene phenocrysts also have higher La/Yb values than ible elements than phenocrysts that crystallized from transitional and do other Leg 144 transitional and alkalic lavas (Fig. 6). Again, it is alkalic series lavas. The separation between basanitic phenocrysts and difficult to evaluate the significance of the high trace element abun- alkalic series phenocrysts occurs at approximately ten times chondritic dances and high La/Yb in the Wodejebato phenocrysts, but differences values for the light rare earth elements (LREE) (Fig. 5). This, in part, in pressure of crystallization, degree of partial melting, and mantle reflects the lower degree of partial melting commonly postulated for source chemistry are the most likely explanations. basanitic series lavas when compared to alkalic series lavas. The ankaramitic lavas recovered from Site 874 were classified as A clear distinction exists between the Limalok and MIT guyot a single igneous unit (a single flow) by D.M. Christie and J. J. Dieu on basanitic phenocrysts. The Limalok Guyot phenocrysts have higher board ship (Premoli Silva, Haggerty, Rack, et al., 1993). There was LREE and lower heavy rare earth elements (HREE) than the MIT no evidence (i.e., modal changes, brecciated flow tops) for indicating Guyot basanitic phenocrysts (cf. Fig. 5A, D). This difference is also more than one flow. However, it is clear from the ion microprobe data illustrated on a plot of La/Yb vs. Ti (Fig. 6), which shows that the (cf. Samples 144-874B-24R-1,12-16 cm, and 144-874B-24R-4,11- Limalok phenocrysts have higher La/Yb values. Assuming that all the 15 cm, in Table 5; see also Figs. 5C, 6) and from the oxide chemistry basanitic lavas were derived from equivalent degrees of partial melt, (Gee, this volume) that at least two flows were recovered. The upper and that all microphenocrysts were crystallized during or after eruption of these lavas, represented by Sample 144-874B-24R-1, 12-16 cm, (1 bar of pressure), differences in pressure of crystallization and of was apparently derived from a less enriched source than the sources degree of partial melt cannot explain this difference. Furthermore, this of the lower lava and the Lo-En, MIT, and Takuyo-Daisan transitional difference is clearly not caused by fractionation because the Mg#s of and alkalic lavas. the phenocrysts in these lavas are essentially equivalent. Therefore, it appears that the source of the Limalok Guyot basanitic lavas was more Microphenocrysts enriched than the source of the MIT Guyot basanitic lavas. This is Four clinopyroxene microphenocrysts were analyzed for REE and verified by whole-rock chemistry (Christie et al., this volume) and by other trace elements to understand better the fractionation processes isotopic ratios (Koppers et al., this volume). Specifically, the Limalok affecting the liquid compositions of Leg 144 lavas before and during Guyot basanitic lavas have higher Nb/Zr ratios and higher 87Sr/86Sr. eruption (Table 5). All four are well-formed equant crystals less than 2. The transitional and alkalic series phenocrysts from Lo-En, 50 µm in diameter. On a diagram of La vs. Yb (Fig. 7), the analyses MIT (thin section Sample 144-878A-23R-1, 6-8 cm), and Takuyo- from Sample 144-878A-89R-2,42-43 cm, provide an integrated pic- Daisan guyots have very similar abundances of REE and other trace ture of the processes that led to the elevated trace element concentra-
523 J.J. DIEU
1000 ß 1000 Limalok Guyot Lo-En Guyot
144-872B-5R-3, 114-115 cm CPX A * 100 α 100 144-872B-9R-5, 107-109 cm CPX A E E 144-872B-9R-5, 107-109 cm CPX B
10 10
-ö 144-871C-35R-1, 37-39 cm CPX B •O 144-871C-38R-1, 29-32 cm CPX A • O- 144-871C-38R-1, 29-32 cm CPX B
La Ce Sr Nd Zr Sm Eu Ti Dy Y Er Yb La Ce Sr Nd Zr Sm Eu Ti Dy Y Er Yb
1000 ; I I I Q1000 ' —π— 144-873 A-18R-1, 30-33 •m CPX A : 144-873A-18R-1, 94-95 ( m CPX A —ö 144-878A-83R-1, 6-8 cm CPX A 144-874B-24R-1, 12-16 < m CPX A 0 144-878A-83R-1, 6-8 cm CPX B —-o—o • 144-874B-24R-1, 12-16 ( m CPX B —Δ— 144-874B-24R-1, 12-16 < m CPX c -•O---- 144-878A-89R-2, 42-43 cm CPX A ---ffl--- 144-874B-24R-4, 11-15 ( m CPX A —ù 144-878A-89R-2, 42-43 cm CPX B 100 -- 144-874B-24R-4, 11-15 ( m CPX B - ra 100 --EB--- 144-878A-89R-2, 42-43 cm CPX D — y— 144-876A-17R-1, 56-61 ( m CPX A : —a— 144-877A-20R-5, 22-24 < m CPX A : 144-877A-20R-5, 22-24 ( m CPX B •
10 - - 10
Hk *"—-– y -O' o Wodejebato Guyot b" MIT Guyot \* -a
La Ce Sr Nd Zr Sm Eu Ti Dy Y Er Yb La Ce Sr Nd Zr Sm Eu Ti Dy Y Er Yb
E 1000 Takuyo-Daisan Guyot
100 144-879A-22R-1, 133-135 cm CPX A 144-879A-22R-1, 133-135 cm CPX B
10
La Ce Sr Nd Zr Sm Eu Ti Dy Y Er Yb
Figure 5. Ion microprobe analyses of clinopyroxene phenocrysts. Chondrite-normalizing values from Evensen et al. (1978). Elements are plotted with increasing incompatibility to the left.
tions in the microphenocrysts. As crystallization progressed from the roxene and olivine were fractionating from the liquid. Roughly 50% formation of phenocryst rims to microphenocryst cores to microphe- total fractionation between crystallization of the rim and the micro- nocryst rims, there was a large increase in the abundances of both La phenocryst is indicated. Ion microprobe analysis was not performed and Yb, reflecting extensive fractionation of the liquid as the lava on a phenocryst from igneous Unit 1 of Site 871. The composition of erupted and chilled. Within the limits of variability of clinopyroxene a microphenocryst from this lava is plotted (Sample 144-871C-35R- chemistry, fractionation of clinopyroxene or olivine cannot be distin- 1,37-39 cm). guished on Figure 7, but petrographic evidence suggests that sub- Large oscillations in trace element concentrations in single pheno- equal proportions of these two phases were crystallizing as the lava crysts have been observed: (1) Shimizu and le Roex (1986) attributed chilled. Using reasonable partition coefficients (Table 6), the calcu- oscillations in augite phenocrysts in alkaline basalts from Gough lated extent of fractionation from crystallization of the phenocryst Island as arising from repeated magma injections into the chamber; rim to crystallization of the microphenocryst rim is roughly 65%. and (2) Shimizu (1990) suggested that oscillations in augite pheno- (Note: There is no reason, beyond the plausibility of this explanation crysts in an alkaline basalt from Papua New Guinea may be caused for the trend observed in Fig. 7, to indicate that the rim of the phe- by nonequilibrium kinetics of reaction-transport of the liquid around nocryst crystallized before the crystallization of the micropheno- the growing crystals. These oscillations may be common in pheno- crysts.) The tie line on Figure 7 from phenocryst rim to micropheno- crysts (Shimizu, 1990), and, therefore, the core-to-rim decrease in cryst for Sample 144-871C-38R-1, 29-32 cm, closely parallels the both La and Yb in Sample 144-871C-38R-1, 29-32 cm, and in La in tie-line sequence for Sample 144-878A-89R-2, 42-43 cm. Again, Sample 144-878A-89R-2, 42-43 cm, is probably not unusual but petrographic evidence suggests that subequal proportions of clinopy- merely the result of too few analyses of a complex system.
524 MINERAL COMPOSITIONS OF LAVAS AND XENOLITHS
25 Micro- Micro- rim pheno- 20- cryst Micro- m pheno -cryst Clinopyroxene
Rim Micro- Olivine core Core im
5000 10000 15000 20000 25000 30000 12 3 4 Ti (ppm) Yb (ppm) D Hole 871C ffl 144-871C-35R-1, 37-39 cm CPX A O Hole 872B o Hole 873A Φ 144-871C-38R-1, 29-32 cm CPX A.B.C• Δ 144-874B-24R-1, 12-16 cm O 144-878A-89R-2, 42-43 cm CPX A,B,C-,D ffl 144-874B-24R-4, 11-15 cm Figure 7. La vs. Yb for clinopyroxene phenocrysts and microphenocrysts. Hole 876A Clinopyroxene- and olivine-controlled fractionation trends are shown. Analy- • Hole 877A ses labeled "core" and "rim" are phenocryst cores and rims; "micro-core" and "micro-rim" are microphenocryst cores and rims; "microphenocryst" repre- V Hole 878A alkalic series sents analyses of cores of very small grains. Tie lines connect analyses of the B Hole 878A basanitic series same lava. Φ Hole 879A Comparisons of Figures 8A and 9A with Figures 8B and 9B, Figure 6. Diagram of La/Yb vs. Ti (ppm) for ion microprobe analyses of respectively, show that the clinopyroxene in the Type 1 xenolith from phenocrysts. Dashed lines connect analyses of basanitic lavas. Solid lines MIT Guyot has higher Mg#s, somewhat lower TiO2 (although clino- connect analyses from the two Hole 874B ankaramitic samples. pyroxene compositions from each guyot have so little Ti that analyti- cal error is more significant than differences between the clinopyrox- ULTRAMAFIC XENOLITHS ene compositions), and significantly lower tetrahedral Al (also overall
lower A12O3; see Table 3) than the clinopyroxene in the Type 1 Ultramafic xenoliths recovered from Sites 871 and 878 (Limalok xenoliths from Limalok Guyot. This indicates that the Type 1 xenolith and MIT guyots, respectively) include two types: Type 1 xenoliths from MIT Guyot represents more severely melted material than the interpreted to be samples of mantle residue and Type 2 xenoliths Type 1 xenoliths from Limalok Guyot. interpreted to be samples of cumulate material (after Frey and Prinz, A goal inherent in the analysis of Type 2 xenoliths is to charac- 1978). Major element analyses by electron microprobe of unaltered terize the liquids that differentiated, forming the cumulate material. mineral phases are presented in Tables 2 and 3. Trace element analy- Evaluation of mineral data from cumulates is very difficult because ses by ion microprobe of representative clinopyroxene grains are depth of crystallization and rate of cooling can drastically affect the presented in Table 5. major element chemistry of all phases. For example, the Salt Lake Crater cumulate xenoliths (Hawaiian hotspot; Sen, 1988) and the Ua Major Element Chemistry Huka cumulate xenoliths (Marquesas hotspot; J.J. Dieu and J.H. Clinopyroxene Natland, unpubl. data, 1993) probably crystallized from very similar liquids (i.e., many lavas of the Honolulu Volcanics [the magmas The clearest distinction between clinopyroxene compositions of known to have formed the Salt Lake Crater pyroxenites] [Sen, 1988])
Type 1 and 2 xenoliths is in their TiO2 contents (Figs. 8, 9). Type 1 and are essentially identical to basanitic and alkalic series lavas on Ua materials have experienced severe melt extraction that has depleted Huka Island (thought to be the magmas that crystallized to form the them in Fe, Ti, and other "incompatible" elements (see discussion of Ua Huka cumulates [J.J. Dieu and J.H. Natland, unpubl. data, 1993]) Type 1 xenoliths under "Trace Element Chemistry" this section). Cli- (cf. Clague and Frey, 1982; Diraison, 1991). Yet there are very signi- nopyroxene grains in Type 1 xenoliths from both Limalok and MIT ficant differences in the clinopyroxene compositions of each xenolith guyots have Mg#s that cluster in a small range (0.895-0.93) and TiO2 locality. Salt Lake Crater clinopyroxene grains have lower TiO2 at an values of < 0.20%. This is typical of mantle residue materials sampled equivalent Mg# (Fig. 8) and only overlap the lower half of the Al(z) by hotspot volcanism (e.g., Sen, 1988; Hauri et al., 1993; J.J. Dieu and range (Fig. 9) observed for Ua Huka clinopyroxene grains. Aluminum J.H. Natland, unpubl. data, 1993). Type 1 analyses from both guyots preferentially moves out of the tetrahedral sites and into the octahe- plot near the field of Samoan mantle residue drawn on Figure 8. dral sites in clinopyroxene under high pressures during solid-solid
525 Table 5. Representative ion microprobe analyses.
Hole: 871C 871C 871C 871C 871C 871C 871C 871C 871C 872B 872B 872B 873A Core, section: 35R-1 35R-1 35R-1 35R-1 35R-1 35R-1 38R-1 38R-1 38R-1 5R-3 9R-5 9R-5 18R-1 Interval (cm): 10-12 10-12 20-23A 20-23A 37-39 37-39 29-32 29-32 29-32 114-115 107-109 107-109 30-33 Description: Alkalic Alkalic Mantle Mantle Alkalic Ground- Core of Rim of Ground- Core of Core of large, Core of large, Core of small cumulate cumulate residue residue cumulate mass 1.0-cm 1.0-cm mass 6-mm, spongy spongy phenocryst; xenolith xenolith xenolith xenolith xenolith xenocryst; xenocryst; round, xenocryst; xenocryst; see CPX 3, 4 grain grain grain grain grain CPXB CPXA spongy see CPX 11 see CPX 11 analyses xenocryst analysis analysis Analysis: CPXA CPXB CPXA CPXB CPXA CPXB CPXA CPXB CPXC CPXA CPXA CPXB CPXA Analysis no.: 1 2 3 4 5 6 7 8 9 10 11 12 13
Element (ppm): Ti 6640 7240 700 720 22790 10410 26320 10030 22210 4680 6260 6080 12110 V 300 310 230 370 370 450 210 460 310 320 310 490 Cr 1010 1010 5300 5560 2790 97 190 1720 107 5480 5190 5620 1940 Sr 194 200 210 300 124 223 145 88 135 44.5 48.2 47.5 63.0 Y 15.3 15.6 7.9 8.4 19.9 19.8 28.9 13.3 35.5 9.3 9.9 10.4 21.8 Zr 140 150 1.3 5.7 240 120 230 140 450 28.1 36.0 33.8 117 La 31.1 9.5 37.7 41.0 57.7 18.2 9.37 8.04 15.6 0.82 1.31 1.39 2.33 Ce 73.9 24.4 48.8 65.2 125.5 48.1 31.5 24.9 40.3 2.72 4.51 4.42 8.06 Nd 36.4 19.7 6.0 13.7 65.6 29.3 31.0 23.8 33.2 3.91 5.37 6.06 10.8 Sm 7.52 4.04 0.49 1.30 14.1 6.63 9.31 7.23 8.65 1.50 1.62 1.87 3.38 Eu 2.31 1.42 0.23 0.51 4.72 2.73 3.13 2.27 2.81 0.61 0.79 0.70 1.56 Dy 4.08 2.29 1.17 1.41 8.07 4.47 6.72 5.22 6.32 1.52 1.94 1.71 3.42 Kr 2.36 1.26 0.95 1.10 4.39 2.24 3.41 2.69 3.04 0.60 0.84 0.90 1.50 Yh [.53 0.83 0.72 0.81 2.87 1.69 2.51 2.09 2.34 0.61 0.64 0.74 1.31
Hole: 874B 874B 874B 874B 874B 876A 877A 877A 878A 878A 878A Core, section: 24R-1 24R-1 24R-1 24R-4 24R-4 17R-1 20R-5 20R-5 23R-1 23R-1 81R-2 Interval (cm): 12-16 12-16 12-16 11-15 11-15 56-61 22-24 22-24 6-8 6-8 37-40 Description: Core of Near edge of Core of 1.0-cm, L-shaped, 5.0-mm resorbed Small crystal Core of Core of Rim of Core of Tholeiitic small 1.0-cm, euhedral euhedral, spongy skeletal xenocryst with in clot; large large resorbed resorbed cumulate phenocryst spongy phenocryst; phenocryst; phenocryst plagioclase see CPX 3 phenocryst phenocryst xenocryst; xenocryst; xenolith CPXC CPXB inclusions analyses CPXB CPXA grain Analysis: CPXA CPXB CPXC CPXA CPXB CPXA CPXA CPXB CPXA CPXB CPXA Analysis no.: 15 16 17 18 19 20 21 22 23 24 25
Element (ppm): Ti 8230 2910 3660 7940 10430 6810 4680 5960 8900 7190 203 V 380 200 230 430 550 330 300 360 330 270 79.5 Cr 6270 6570 6300 5500 2690 2710 1890 3850 3250 2160 700 Sr 43.7 26.0 34.6 45.7 41.2 53.2 31.1 34.3 42.6 41.5 46.7 Y 13.5 6.0 7.0 14.1 20.8 11.1 8.2 11.0 14.1 12.1 0.6 Zr 52.6 11.1 16.8 65.4 71.2 53.0 23.8 32.7 46.3 30.4 12.2 La 0.87 0.41 0.47 3.01 3.20 3.29 2.62 2.06 1.34 1.87 0.31 Ce 3.26 1.29 1.50 10.7 11.4 10.7 8.76 6.40 5.17 6.11 0.74 Nd 4.25 2.31 1.99 9.17 9.91 8.24 7.14 5.84 5.73 7.00 0.59 Sm 1.60 1.07 0.80 3.03 3.20 3.00 2.99 2.66 2.77 3.04 0.38 Eu 0.61 0.37 0.37 1.09 1.39 0.95 1.11 0.83 1.14 1.10 0.14 Dy 1.81 0.59 0.96 2.80 3.99 2.05 M•O 2.37 2.77 3.30 0.42 Er 0.82 0.25 0.56 1.65 1.74 1.21 1.16 1.10 1.33 1.65 0.49 Yb 0.74 0.39 0.38 1.05 1.37 1.04 1.05 0.97 1.13 1.27 0.43 MINERAL COMPOSITIONS OF LAVAS AND XENOLITHS
equilibration (e.g., this may affect cumulate grains but not pheno- crysts) (Le Bas, 1962). Because silica must diffuse into clinopyrox- ene for this to occur and, thus, the tetrahedral charge deficit is dimin- ON ON « •taiqnooqoi-; i•tcnd — -3- •* Orthopyroxene Unaltered orthopyroxene was analyzed in three xenoliths from < 2 s é thin section Samples 144-871C-35R-1,20-23 cm (Xenolith A), 144- £00 σs <^^ ε g•< << (Sen, 1988). Thus, the presence of orthopyroxene in this one set of 527 J.J. DIEU t.u- 20.0 Φ Type 1 xenolith 144-871C-35R-1 23-25A cm Φ Type 1 xenolith 144-871C-35R-1, 23-25A cm O Type 1 xenolith 144-871C-35R-1 23-25B cm O Type 1 xenolith 144-871C-35R-1, 23-25B cm D Type 2 xenolith 144-871C-35R-1,10-12 cm D Type 2 xenolith 144-871C-35R-1, 10-12 cm Δ Type 2 xenocryst core 144-871C-35R-1 14-16 cm Δ Type 2 xenocryst core 144-871C-35R-1,14-16 cm 3.0- EH Type 2 xenocryst 144-871C-35R-1, 23-25B 15.0- ffl Type 2 xenocryst 144-871C-35R-1, 23-25B O Type 2 xenolith 144-871C-35R-1, 37-39 cm O Type 2 xenolith 144-871C-35R-1,37-39 cm Field of Ua Huka y× Field of Ua Huka cumulate xenoliths/ cumulate xenoliths CM ^ 10.0^ O 2.0- Λ < Alkaline Field of Salt Lake / Crater cumulate ( xenoliths ^v Δ 1.0- 5.0- v< / Nonalkaline Field of Samoan —~^. residual xenoliths J Field of Salt Lake Crater cumulate xenoliths / 0.0 o.o- —i— —l— —I— 1 i o.o 1.0 2.0 3.0 4.0 5.0 0.95 0.90 0.85 0.80 0.75 TiO % Mg/(Mg + Fe) 2 B B 20.0 4.0 Type 1 xenolith 144-878A-81R-2, 37-40 cm D Type 1 xenolith 144-878A-81R-2, 37-40 cm Type 2 xenolith 144-878A-90R-5, 111 -116 cm O Type 2 xenolith 144-878A-90R-5, 111 -116 cm Type 2 xenolith 144-878A-90R-6, 63-64 cm O Type 2 xenolith 144-878A-90R-6, 63-64 cm Type 2 xenolith 144-878A-91R-5, 47-54 cm Δ Type 2 xenolith 144-878A-91R -5, 47-54 cm 15.0- 3.0- Field of Ua Huka Field of Ua Huka cumulate xenoliths cumulate xenoliths 10.0- Alkaline Field of Salt Lake Crater cumulate xenoliths 5.0- 1.0- Nonalkaline Field of Samoan residual xenoliths o.o • 0.0 o.o 1.0 2.0 3.0 4.0 5.0 0.95 0.90 0.85 0.80 0.75 TiO2 % Mg/(Mg + Fe) Figure 9. Al(z) vs. TiO2 of clinopyroxene grains in xenoliths. A. Limalok Guyot xenoliths. B. MIT Guyot xenoliths. Boundary lines from Le Bas (1-962). Fields Figure 8. TiO2 vs. Mg/(Mg+Fe) of clinopyroxene grains in xenoliths. A. Limalok Guyot xenoliths. B. MIT Guyot xenoliths. Fields drawn from data in from sources listed in Figure 8 caption. Sen (1988) for Salt Lake Crater xenoliths and from unpublished data of JJ. Dieu and J.H. Natland (1993) for Ua Huka xenoliths. be pieces of a mantle source that was partially melted before being sampled as a xenolith. Such xenoliths erupted at an oceanic hotspot are often samples of oceanic lithosphere (e.g., Sen, 1988; K.A. Farley, alkalic cumulate xenoliths does not undermine the hypothesis that the unpubl. data, 1994). MIT Guyot cumulates were derived from tholeiitic liquids. This extreme amount of partial melting should be reflected in low incompatible trace element abundances. On an extended REE dia- Trace Element Chemistry gram (Fig. 10A), data for a clinopyroxene in a Type 1 xenolith from Tahiti display the low abundances and pattern representative of man- Type 1 Xenoliths tle residue. However, xenoliths from the Southern Hemisphere oce- As previously discussed in the "Major Element Chemistry" sec- anic hotspots seldom display these expected patterns. In many cases, tion (this chapter), Type 1 (mantle residue) xenoliths are presumed to some process subsequent to melting has elevated the abundances of MINERAL COMPOSITIONS OF LAVAS AND XENOLITHS Table 6. Constants used in percolation model. Table 7. Comparative orthopyroxene analyses. Equilibrium SAV 1-5 A Partition melt values OPX/1 33-H coefficients CPX OPX OL SP (CN) SiO2 52.79 54.04 La 0.045 0.0015 0.0006 0.0003 3500 TiO2 0.00 0.28 Ce 0.075 0.0026 0.003 0.0005 1600 A12O, 1.35 4.15 Sr 0.095 0.001 0.0028 0.0006 570 Cr2O3 0.67 0.00 Nd 0.1 0.009 0.003 0.0008 330 FeO 6.52 11.07 Zr 0.15 0.015 0.003 0.003 20 MnO 0.19 ND Sin 0.2 0.012 0.003 0.0009 80 MgO 38.40 29.50 Eu 0.2 0.012 0.003 0.0009 60 NiO 0.00 ND Ti 0.34 0.1 0.01 0.005 5 CaO 0.90 0.97 Dy 0.35 0.054 0.006 0.0015 24 Na2O 0.00 0.00 Y 0.37 0.0481 0.007 0.0025 3.4 Total 100.82 100.01 Er 0.37 0.076 0.008 0.003 14 Yb 0.37 0.11 0.009 0.0045 8 En 89.94 80.99 13 13 Fs 8.53 17.08 Diffusion LOO" l.OO~ LOO"12 LOO"13 Wo 1.53 1.93 coefficient (cm/a) Modal 1 15 82 2 abundance (%) Note: SAV 1-5A OPX/1 = Samoan mantle residue (data from J.J. Dieu, unpubl. data, Grain size (cm) 0.2 1.5 2.5 0.1 1993). 33-H = Pyroxenite cumulate (data from Sen, 1988). ND = no data. Notes: Porosity = 1%; percolation distance = 100 m; percolation speed = 1 cm/a. The equilibrium melt values are chondrite normalized (CN). Partition coefficients are represents equivalently melted mantle. The initial composition of the within the ranges specified by Arth (1976). percolating melt is plotted on Figure 10H. The early percolation stages of the model are a fairly good reproduction of the data from Sample 144-871C-35R-1, 23-25 cm (Xenolith A). The clinopyrox- the large ion lithophile elements (LILE) without obviously modify- ene pattern that would represent final equilibration with the percolat- ing the modal abundances and major element chemistry of the resi- ing liquid is plotted on Figure 11C. due. A process with this signature is called "cryptic metasomatism" The second type of Type 1 clinopyroxene pattern (Fig. 10E, F) is (Dawson, 1984; Zindler and Jagoutz, 1988). Extended REE patterns typically quite smooth and slightly convex-upward; it also displays like those plotted on Figure 10 (B, D, E, F) from Limalok Guyot, very high abundances of REE [e.g., 40-100 La(n)] with strong nega- Savai'i Island (Samoa), MIT Guyot, and Tubuai Island (Cook- tive HFSE anomalies. Different xenoliths from a locality on Savai'i, Australs), respectively, are typical of Type 1 clinopyroxene patterns Western Samoa, contain clinopyroxene with both types of patterns from the Samoan, Marquesan, Cook-Austral, and Society hotspot (J.J. Dieu and J.H. Natland, unpubl. data, 1993). Xenoliths with chains (Hauri et al., 1993; J.J. Dieu and J.H. Natland, unpubl. data, clinopyroxene that have this type of pattern were also cryptically 1993) and from the Kerguelen plume (Hassler et al., 1993). metasomatized (i.e., no hydrous phases were added and there are no Two distinct extended REE patterns of Type 1 clinopyroxene are observable changes in major element chemistry). These two lines of observed from the Southern Hemisphere hotspots. The first type of evidence—the occurrence of both patterns from a single xenolith pattern is "spoon-shaped" (Sen et al., 1993) with flat and low (less locality, sampling otherwise identical material, and both patterns cre- than ten times chondrite) HREE abundances and strong upward flex- ated by a process best described as cryptic metasomatism—suggest ure with increasing incompatibility in the LREE. Middle rare earth that this second type of pattern is also created by percolation. In this elements (MREE) may or may not have begun to exhibit increasing latter case, percolation occurred for a sufficient time period for the abundances. Often the high field strength elements (HFSE), such as less incompatible elements to arrive and equilibrate before the mate- Zr and Ti, are at or so close to their post-melting abundances that they rial was sampled as a xenolith. Basically, these clinopyroxene grains create strong negative anomalies in the chondrite-normalized pat- were in equilibrium with a very enriched melt; percolation is invoked terns. The clinopyroxene analyses plotted on Figure 10 (B, D) are as a reasonable, but not required, explanation. typical of this type of pattern. Melts in equilibrium with the clinopyroxene analyses from MIT This type of pattern can be modeled as developing from a low- Guyot Type 1 xenoliths (Fig. 10E) were calculated using the partition porosity melt percolation process that chromatographically fraction- coefficients in Table 6 and are plotted, chondrite-normalized, on ates elements from one another (Navon and Stolper, 1987; Bodinier Figure 10G. These calculated melts and the estimated melt composi- et al., 1990). As with cation exchange resin in a laboratory column, tion used in the modeling of the Limalok Guyot Type 1 clinopyroxene the basic process is one of speed of transport dependent on an ele- data show unusually high REE abundances and slight Zr and large Ti ment^ proportion of time spent bonded to the mantle residue vs. time negative anomalies. These characteristics preclude the metasomatiz- spent moving with the melt. The more incompatible elements move ing melts being basaltic liquids or their differentiates. Actual erupted at a speed closer to that of the melt front because they spend less time liquids that most closely resemble these percolating melts are car- in mantle minerals. Thus, the LREE will arrive at a certain point (later bonatites (cf. Fig. 10G, H). This was also the conclusion of Hauri et sampled as a xenolith) above the percolation starting point before the al. (1993) for Samoan and Cook-Austral xenoliths and Hassler et al. MREE and HREE. Diffusion of these early-arriving elements into the (1993) for Kerguelen xenoliths. residual mantle will cryptically metasomatize the material. For this fragment of the mantle to be sampled, a much larger melt fraction en route to the surface must mechanically fracture into the area and Type 2 Xenoliths entrain samples of a material that has interacted only with the leading Clinopyroxene grains from two Type 2 xenoliths, one from Li- edges of a percolating melt. malok Guyot and one from MIT Guyot, were analyzed by ion micro- The model presented in Figure 10C was calculated using the probe for rare earth and other trace element concentrations. These Fortran percolation program of Bodinier et al. (1990), as modified for data are presented in Table 7. Hereafter in this section, these xenoliths Macintosh by E. Takazawa (unpubl. data, 1992). The partition coef- are referred to as the "Limalok xenolith" and the "MIT xenolith." ficients and other constants used for this model are presented in Table Figure 11A is a chondrite-normalized extended REE diagram. 6. The alteration of the xenolith in thin section Sample 144-871C- Two analyses from each of the Limalok and MIT xenoliths are plot- 35R-1, 23-25 cm (Xenolith A), precluded modal analysis and grain- ted. There are some small anomalies, both positive and negative, size estimation. These were taken from a Tahitian xenolith that I think in the trace element abundances relative to the REE abundances. 529 J.J. DIEU π 1 1 1 1 I 1 1 1 r ß 1000 Unmetasomatized 144-871C-35R-1, 23-25A cm CPX A Tahitiaπ mantle residue 144-871C-35R-1, 23-25A cm CPX B 10 2 100 10 0.1 TAH 6-3 CPX 1 TAH 6-3 CPX 2 0.01 0.1 La Ce Sr Nd Zr Sm Eu Ti Dy Y Er Yb La Ce Sr Nd Zr Sm Eu Ti Dy Y Er Yb 100 Percolation Model for 144-871C-35R-1, Samoan mantle 10-12 cm CPX A and B residue 100 - 10 0.1 0.01 La Ce Sr Nd Zr Sm Eu Ti Dy Y Er Yb La Ce Sr Nd Zr Sm Eu Ti Dy Y 1000 1000 Tubuai mantle residue 2 100 100 10 10 D 144-878A-90R-6, 63-64 cm CPX A O 144-878A-90R-6, 63-64 cm CPX B TBA 4-11 CPX 1 •―O— 144-878A-91R-1, 88-89 cm CPX A TBA 1-9 CPX 2 0.1 0.1 La Ce Sr Nd Zr Sm Eu Ti Dy Y Er Yb La Ce Sr Nd Zr Sm Eu Ti Dy Y Er Yb Figure 10. Chondrite-normalizing values from Evensen et al. (1978). A. Ion microprobe data of clinopyroxene in a Tahitian Type 1 xenolith. B. Ion microprobe data of Limalok Guyot Type 1 clinopyroxene. C. Modeled clinopyroxene patterns. Steps at 500, 1500, 2500, and 3500 yr. Constants in Table 6. D. Ion microprobe data of Samoan Type 1 clinopyroxene (from J.J. Dieu and J.H. Natland, unpubl. data, 1993). E. Ion microprobe data of MIT Guyot Type 1 clinopyroxene. F. Ion microprobe data of Cook-Austral Type 1 clinopyroxene (from Hauri et al., 1993). G. Melts calculated to be in equilibrium with MIT Guyot Type 1 clinopyroxene and the melt used for the percolation model in Figure 1 IB. H. Shaded field encloses data for erupted carbonatites reported by Nelson et al. (1988). Small, random-direction anomalies most likely represent heteroge- differentiated to form the cumulate material sampled by the xenoliths. neities within the clinopyroxene grains. Rare earth and trace elements The Limalok cumulate-forming liquids were much more enriched in are analyzed at separate times on distinct spots about 10 µm apart, and LREE, Sr, Zr and, by inference, other highly incompatible elements. heterogeneity in many of the crystals can be measured "downhole" This agrees with the evidence of the major element chemistry of the on a submicron scale during an analysis. clinopyroxene, which suggests crystallization from alkalic liquids. The clinopyroxene grains in the Limalok xenolith contain greater Figure 11 (B, C, D) present sets of clinopyroxene data from Type abundances of the LREE and of Sr and Zr than do the clinopyroxene 2 xenoliths from Ua Huka Island in the Marquesas Chain, Tahiti grains in the MIT xenolith, and, overall, REE patterns of the Limalok Island in the Society Chain, and the Island of Hawaii, respectively. xenolith clinopyroxene grains are steeper (i.e., have greater La/Yb) Each of these suites of xenoliths has been interpreted as samples of than those of the MIT xenolith clinopyroxene grains. These differ- alkalic cumulates (Chen et al., 1992; J.J. Dieu and J.H. Natland, ences reflect a fundamental difference in the types of liquids that unpubl. data, 1993). It is clear from the total range of alkalic cumulate MINERAL COMPOSITIONS OF LAVAS AND XENOLITHS 10000 H 10000 f- 1000 1000 r ε 100 - 100 r 10 A r f A " Melt in equilibrium with 144-878A-90R-6, 63-64 cm CPX A O Melt in equilibrium with 144-878A-90R-6, 63-64 cm CPX B -•--O- -• Melt in equilibrium with 144-878A-91R-1, 88-89 cm CPX A —Δ— Melt used in calculation of model for 144-871C-35R-1, 10-12 cm 0.1 La Ce Sr Nd Zr Sm Eu Ti Dy Y Er Yb La Ce Sr Nd Zr Sm Eu Ti Dy Er Yb Figure 10 (continued). clinopyroxene patterns presented in Figure 11 (B, C, D) that identifi- rock chemistry and petrography of Limalok and some MIT lavas as cation of the MIT xenolith as a tholeiitic cumulate on the basis of this basanitic are strongly supported by mineral chemistry. evidence is not possible. Presumably, clinopyroxene grains in tholei- High La/Yb ratios of phenocrysts in the Limalok Guyot basanitic itic cumulates also show a very wide range of rare earth and trace lavas indicate that these were created from a more enriched source elements, but the data are not yet available for comparison with the than were the MIT Guyot basanitic lavas. Whole-rock trace element MIT xenolith data. data and isotopic data support this conclusion. The clinopyroxene grain analyzed from Sample 144-871C-35R- The La/Yb ratios and chondrite-normalized abundances of rare 1, 10-12 cm (CPX A), contains much greater abundances of LREE earth and trace elements in phenocrysts from alkalic series lavas from than I have previously observed in cumulate clinopyroxene grains Lo-En, MIT, and Takuyo-Daisan guyots are similar, despite apparent from the Samoan, Society, or Marquesan chains. Because the identi- differences in source chemistry (Koppers et al., this volume), in the fication of this xenolith as Type 2 is firmly established on arguments (probable) degrees of partial melting from which each lava suite was of petrography (no orthopyroxene) and major element chemistry, this derived, and in the degrees of differentiation. This coincidence illus- analysis provides a new upper limit to the range of LREE abundances trates the difficulties inherent in working with phenocryst chemistry. observed in clinopyroxene from alkalic cumulates. The range, as The La/Yb ratios and chondrite-normalized abundances of rare presented in Figure 11, shows a variance in La abundances across 2.5 earth and trace elements in phenocrysts indicate that the ankaramite orders of magnitude. Mafic alkalic series lavas that erupt at hotspots recovered from Hole 874B consists of at least two flows apparently do not exhibit a range anywhere near this great; they vary, perhaps at derived from sources of distinctly different enrichment. The less most, by a factor of 3 in La concentration. Obviously, a process other enriched source was less enriched than the source materials of the than simple equilibrium phenocryst growth is controlling the incom- transitional and alkalic series lavas from Lo-En, MIT, and Takuyo- patible trace element abundances in cumulate clinopyroxenes. Lang- Daisan guyots. The phenocrysts of the ankaramitic lava with more muir (1989) hypothesized and modeled a process he called "in situ enriched signatures have higher abundances of trace elements and fractional crystallization," which described the effects of fractiona- higher La/Yb than do phenocrysts in the Lo-En, MIT, and Takuyo- tion on an interstitial liquid as post-cumulus growth occurs during Daisan transitional or alkalic series lavas for any of several reasons. cumulate solidification. Stated simply, the results of this type of Incompatible element abundances in phenocryst rims and micro- fractionation are liquids with only slightly modified major element phenocryst cores and rims show that fractionation of these liquids was chemistry but with very elevated abundances of incompatible ele- extensive (50%-65%) just before and during eruption. ments. These pockets of interstitial liquid may crystallize clinopy- Major and trace element chemistry of clinopyroxene from Type 2 roxene with Mg#s similar to the original cumulus cores, but with high xenoliths from Limalok Guyot indicate that these are most likely and variable abundances of LREE and other incompatible elements. cumulate materials that fractionated from alkalic liquids. Presumably, Cumulates formed in shallow, rapidly cooling magma chambers ap- given the high abundances of rare earth and trace elements, these pear to preserve these heterogeneities on a grain- or xenolith-length alkalic liquids were related to earlier volcanism on Limalok Guyot scale because subsolidus diffusion of incompatible elements in clino- (before the eruption of the basanitic-series that entrained the xenolith) pyroxene is much too slow to be effective at erasing them (J.J. Dieu and were not mid-ocean-ridge related. and J.H. Natland, unpubl. data, 1993). Major element chemistry of clinopyroxene and the presence of orthopyroxene, coupled with an apparent lack of plagioclase, indicate CONCLUSIONS that the MIT Guyot Type 2 xenoliths are cumulates of tholeiitic liquids with hotspot characteristics rather than abyssal characteristics. This is Lavas recovered from Wodejebato Guyot, Sites 874 and 877, and the single indication from all drilling results of Leg 144 that these igneous Unit 15 of Site 878 on MIT Guyot are relatively primitive volcanic edifices may be built largely of tholeiitic magma and that compositions. The liquids had Mg#s that ranged from 0.60 to 0.65, transitional, alkalic, and basanitic series lavas form only a carapace. based on calculations of olivine chemistry and comparisons of clino- The high abundances of incompatible elements in the clinopyrox- pyroxene and plagioclase chemistry. ene grains in Type 1 xenoliths from MIT and Limalok guyots suggest Lavas from Limalok and MIT guyots identified as basanitic con- that cryptic metasomatism occurred. The extended REE patterns from tain clinopyroxene phenocrysts and groundmass grains with rather Limalok Guyot Type 1 clinopyroxene grains can be modeled as devel- remarkable compositions characterized by high TiO2 (usually > 2%; oping from diffusion of the early-arriving elements in a percolating Fig. 1), high A12O3 (up to 12%; Table 3), and high tetrahedral alumi- carbonatite-like liquid. Sampling of these xenoliths occurred before num (Al(z) > 10). These characteristics were not observed in clinopy- equilibration with the later-arriving elements could occur. The ex- roxene from the alkalic series lavas. Thus, identifications by whole- tended REE patterns from MIT Guyot Type 1 clinopyroxene grains 531 J.J. DIEU 1000 B 1000 D 144-871C-35R-1, 10-12 cm CPX A U UAH 298-9 CPX 3 0 144-871C-35R-1, 10-12 cm CPX B •~O• — UAH 298-8 CPX 4 ...-O-•-• 144-878A-90R-5, 111-116 cm CPX A —-Δ—- UAH 298-3 CPX 3 Δ 144-878A-90R-5, 111-116 cm CPX B w 100 100 O UAH 298-5 CPX 1 10 10 La Ce Sr Nd Zr Sm Eu Ti Dy Y Er Yb La Ce Sr Nd Zr Sm Eu Ti Dy Y Er Yb C 1000 I 1 1 D1000 P 85T95-5B CPX A 85T95-7 CPX A 0 85T95-5B CPX B 85T95-7 CPX B -•--O---- 85T95-5B CPX C 85T95-7 CPX C rö 100 r ù 8ST9S-SB CPX D 10 La Ce Sr Nd Zr Sm Eu Ti Dy Y Er Yb Ce Nd Figure 11. Analyses of clinopyroxene grains from cumulate xenoliths. A. Limalok and MIT Guyot Type 2 xenoliths. B. Ua Huka Island alkalic cumulate xenoliths (from J.J. Dieu and J.H. Natland, unpubl. data, 1993). C. Tahitian alkalic cumulate xenoliths (from J.J. Dieu and J.H. Natland, unpubl. data, 1993). D. Hualalai (Hawaii) alkalic cumulate xenoliths (from Chen et al., 1992). All patterns plotted in Figure 11 (A-C) are ion microprobe analyses. Patterns plotted in Figure 11D are INAA analyses of mineral separates. appear to be fully equilibrated with a carbonatite-like melt. Xenoliths Chen, C.-Y., Frey, F.A., and Garcia, M.O., 1990. Evolution of alkalic lavas with these types of extended REE patterns have been identified from at Haleakala Volcano, East Maui, Hawaii. Contrib. Mineral. Petrol, the Marquesas and Society chains (J.J. Dieu and J.H. Natland, unpubl. 105:197-218. data, 1993), from the Samoan and Cook-Austral chains (Hauri et al., Clague, D.A., Fisk, M.R., and Bence, A.E., 1980. Mineral chemistry of basalts from Ojin, Nintoku, and Suiko Seamounts Leg 55 DSDP. In Jackson, E.D., 1993), and from the Kerguelen Islands (Hassler et al., 1993). This Koizumi, I., et al., Init. Repts. DSDP, 55: Washington (U.S. Govt. Printing plume process of carbonatitic liquids percolating into previously Office), 607-637. depleted peridotite appears to be a process that is unique to the Clague, D.A., and Frey, F.A., 1982. Petrology and trace element geochemistry enriched Southern Hemisphere mantle sources. It is a process with a of the Honolulu volcanics, Oahu: implications for the oceanic mantle long history in the SOPITA region—at least 128 Ma of history. below Hawaii. /. Petrol, 23:447-504. Davies, H.L., Sun, S.-S., Frey, F.A., Gautier, I., McCulloch, M.T., Price, R.C., REFERENCES* Bassias, Y., Klootwijk, C.T., and Leclaire, L., 1989. Basalt basement from the Kerguelen Plateau and the trail of the DUPAL plume. Contrib. Mineral. Arm, J.G., 1976. Behavior of trace elements during magmatic processes: a Petrol, 103:457^69. summary of theoretical models and their applications. J. Res. U.S. Geol. Dawson, J.B., 1984. Contrasting types of upper-mantle metasomatism? In Surv., 4:41^7. Kornprobst, J. (Ed.), Kimberlites (Vol. 2): The Mantle and Crust/Mantle Bodinier, J.L., Vasseur, G., Vernieres, J., Dupuy, C, and Fabries, J., 1990. Relationships: Amsterdam (Elsevier), 289-294. Mechanisms of mantle metasomatism: geochemical evidence from the Deer, W.A., Howie, R.A., and Zussman, J., 1966. An Introduction to the Lherz Orogenic Peridotite. J. Petrol, 31:597-628. Rock-Forming Minerals: London (Longman Group). Chauvel, C, Hofmann, A.W., and Vidal, P., 1992. HIMU-EM: the French Dick, H.J.B., and Bullen, T, 1984. Chromian spinel as a petrogenetic indicator Polynesian connection. Earth Planet. Sci. Lett., 110:99-109. in abyssal and alpine-type peridotites and spatially associated lavas. Con- Chen, C.-H., Presnall, D.C., and Stern, R.J., 1992. Petrogenesis of ultramafic trib. Mineral. Petrol, 86:54-76. xenoliths from the 1800 Kaupulehu Flow, Hualalai Volcano, Hawaii. J. Diraison, C, 1991. Le volcanisme aerien des Archipels Polynesiens de la Petrol, 33:163-202. Societe, des Marquises et des Australes-Cook—tephrostratigraphie, data- tion isotopique et geochimie comparees—contribution a 1'etude des origi- nes du volcanisme intraplaque du Pacifique Central [these de Doctorat]. Univ. de Bretagne Occidentale. Abbreviations for names of organizations and publication titles in ODP reference lists Evensen, N.M., Hamilton, P.J., and O'Nions, R.K., 1978. Rare-earth abun- follow the style given in Chemical Abstracts Service Source Index (published by dances in chondritic meteorites. Geochim. Cosmochim. Acta., 42:1199- American Chemical Society). 1212. 532 MINERAL COMPOSITIONS OF LAVAS AND XENOLITHS Fodor, R.V., Keil, K., and Bunch, T.E., 1975. Contributions to the mineral Mori, T, and Green, D.H., 1976. Subsolidus equilibria between pyroxenes in chemistry of Hawaiian rocks. IV: Pyroxenes in rocks from Haleakala the CaO-MgO-SiO2 system at high pressures and temperatures. Am. and West Maui volcanoes, Maui, Hawaii. Contrib. Mineral. Petrol., Mineral, 61:616-625. 50:173-195. Navon, O., and Stolper, E., 1987. Geochemical consequences of melt perco- , 1977. Contributions to the mineral chemistry of Hawaiian rocks, lation: the upper mantle as a chromatographic column. J. Geol, 95:285- VI: olivines in rocks from Haleakala and West Maui Volcanoes, Maui, 307. Hawaii. Pac. ScL, 31:299-308. Nelson, D.R., Chivas, A.R., Chappell, B.W., and McCulloch, M.T., 1988. Frey, F.A., and Prinz, M., 1978. Ultramafic inclusions from San Carlos, Geochemical and isotopic systematics in carbonatites and implications Arizona: petrologic and geochemical data bearing on their petrogenesis. for the evolution of ocean-island sources. Geochim. Cosmochim. Acta, Earth Planet. Sci. Lett., 38:129-176. 52:1-17. Frey, F.A., Wise, W.S., Garcia, M.O., West, H., Kwon, S.-T., and Kennedy, A., Premoli Silva, I., Haggerty, J., Rack, E, et al., 1993. Proc. ODP, Init. Repts., 1990. Evolution of Mauna Kea Volcano, Hawaii: petrologic and geochemi- 144: College Station, TX (Ocean Drilling Program). cal constraints on post-shield volcanism. J. Geophys. Res., 95:1271-1300. Roeder, P.L., and Emslie, R.F., 1970. Olivine-liquid equilibrium. Contrib. Garcia, M.O., Frey, F.A., and Grooms, D.G., 1986. Petrology of volcanic rocks Mineral Petrol, 29:275-289. from Kaula Island, Hawaii. Contrib. Mineral. Petrol., 94:461-471. Sen, G., 1988. Petrogenesis of spinel lherzolite and pyroxenite suite xenoliths Garcia, M.O., Ho, R.A., Rhodes, J.M., and Wolfe, E.W., 1989. Petrologic from the Koolau shield, Oahu, Hawaii: implications for petrology of the constraints on rift-zone processes. Bull. Volcanol., 52:81-96. post-eruptive lithosphere beneath Oahu. Contrib. Mineral. Petrol, Gill, J.B., 1981. Minerals and Rocks (Vol. 16): Orogenic Andesites and Plate 100:61-91. Tectonics: Berlin (Springer-Verlag). Sen, G., Frey, F.A., Shimizu, N., and Leeman, W.P., 1993. Evolution of the Gonzales-Marabel, A.-M., 1984. L'ile de Hiva Oa dans les Marquises (Paci- lithosphere beneath Oahu, Hawaii: rare earth element abundances in fique Central): petrologie et evolution volcanologique [Ph.D. thesis]. mantle xenoliths. Earth Planet. Sci. Lett, 119:53-69. L'Universite de Paris-Sud, Centre D'Orsay. Shimizu, N., 1990. The oscillatory trace element zoning of augite phenocrysts. Hack, P.J., Nielsen, R.L., and Johnston, A.D., in press. Experimentally deter- Earth-Sci. Rev., 29:27-37. mined rare earth element and Y partitioning behavior between clinopy- Shimizu, N., and le Roex, A.P., 1986. The chemical zoning of augite pheno- roxene and basaltic liquids at pressures up to 20 kb. Chem. Geol. crysts in alkaline basalts from Gough Island, South Atlantic. J. Volcanol. Hart, S.R., 1984. A large-scale isotope anomaly in the Southern Hemisphere Geotherm. Res., 29:159-188. mantle. Nature, 309:753-757. Staudigel, H., Park, K.-H., Pringle, M.S., Rubenstone, J.L., Smith, W.H.F., and Hassler, D., Shimizu, N., Frey, R, Weis, D., Cottin, J.-Y, and Giret, A., 1993. Zindler, A., 1991. The longevity of the South Pacific isotopic and thermal Geochemistry of peridotite and granulite xenoliths from the Courlet Pen- anomaly. Earth Planet. Sci. Lett., 102:24^4. insula, Northeastern Kerguelen Islands. Eos, 74:632. (Abstract) Takahashi, E., 1980. Melting relations of an alkali-olivine basalt to 30 kbar, Hauri, E.H., Shimizu, N., Dieu, J.J., and Hart, S.R., 1993. Evidence for and their bearing on the origin of alkali basalt magmas. Year Book— hotspot-related carbonatite metasomatism in the oceanic upper mantle. Carnegie Inst. Wash., 79:271-276. Nature, 365:221-227. , 1980. Melting relations of a primitive alkali-olivine basalt. Geol. Herzberg, C.T., 1978. Pyroxene geothermobarometry and geobarometry: ex- Soc. Am. Abstr. Progr., 12:533. (Abstract) perimental and thermodynamic evaluation of some solidus phase relations Vidal, P., Chauvel, C, and Brousse, R., 1984. Large mantle heterogeneities involving pyroxenes in the system CaO-MgO-Al2O3-SiO2. Geochim. Cos- beneath French Polynesia. Nature, 307:536-538. mochim. Acta, 42:945-957. Wells, P.R.A., 1977. Pyroxene thermometry in simple and complex systems. Ho, R.A., and Garcia, M.O., 1988. Origin of differentiated lavas at Kilauea Contrib. Mineral. Petrol, 62:129-139. Volcano, Hawaii: implications from the 1955 eruption. Bull. Volcanol, Wright, E., and White, W.M., 1987. The origin of Samoa: new evidence from 50:35-46. Sr, Nd, and Pb isotopes. Earth Planet. Sci. Lett., 81:151-162. Keil, K., Fodor, R.V., and Bunch, T.E., 1972. Contributions to the mineral Zindler, A., and Jagoutz, E., 1988. Mantle cryptology. Geochim. Cosmochim. chemistry of Hawaiian rocks, II: Feldspars and interstitial material in rocks Acta, 52:319-333. from Haleakala and West Maui volcanoes, Maui, Hawaii. Contrib. Min- eral. Petrol, 37:253-276. Langmuir, C.H., 1989. Geochemical consequences of in situ crystallization. Nature, 340:199-205. Date of initial receipt: 31 January 1994 Le Bas, M.J., 1962. The role of aluminum in igneous clinopyroxenes with Date of acceptance: 19 December 1994 relation to their parentage. Am. J. ScL, 260:267-288. Ms 144SR-029 533